How to Avoid the 7 Stages of Grief When Naming Your Company

This is Part 7/18 in the series “How to Build an Innovative New Product or Company” on the topic of finding a name and URL for your new company

Finally, the entrepreneur thinks, the moment to name my new company—a chance at fun and levity!

An evanescent oasis from the eternal desert of agony that crafting my new enterprise so far has been!

A moment like the final few steps before a mountain summit—a reward so sweet for so little work only because it ruthlessly pilfers from the coffers of the thousands of steps already taken!

Sure, naming my company is a step that could be outsourced but why do all that hard work and then pay someone else for to bask in the glow of the one joyful shining moment?

Then… how quickly this step, too, dissolves into horror and chaos. Like a songwriter who, upon trying every chord progression she can possibly think of, realizes that every song has already been written.

Denial leads to anger … anger leads to bargaining, “Why, oh benevolent Lord, do You grace ‘Apple’ with such a simple name but all that is left for me is JesuitNinny.com? Where have I wronged You? What sacrifice will reverse Your decision?”

Bargaining leads to depression. You started the process of naming your company so bright-eyed and full of hope and youthful energy and now, barely able to get yourself out of bed in the morning, you are willing to prostitute your company in any way to get through the naming process with a shred of a worthy outcome.

Depression finally leads to acceptance. If there were any good names left, someone would have already found them. What you need is not a name that you’ll want tattooed with pride on your arm, but a simple .com, five to nine letters long, easy to spell, and that has some relevance to the direction/values/mission of your company, however remote.

This advice came out of going through the naming process a couple of times. My current company, Laudio, is focused on helping managers in healthcare better recognize and support their staff. The Latin root we found is “laud”, from which derives laudatory and applaud. A quick search on GoDaddy auctions found Laudio.com and then we acquired the trademark shortly afterwards. This whole naming process was done in about an hour—though admittedly after spending months on other approaches.

One nice side effect of this process was it gave us a “naming story”. Prospective customers often ask how we came up with the name and we explain the Latin root. It helps tie our mission and product together in the name, something that actually has quite a bit of value.

My advice:

  1. Find a few English words that represent that heart of your highest, grandest ambitions. For example, for a patient fall prevention product, “safety”, “lives”, “care” are all contenders. As the company grows to include areas beyond fall prevention, it won’t have to rename.
  2. As clichéd as it sounds, go to a website such as Google Translate and look for Latin or Greek roots for those words:
    https://translate.google.com/?tl=el#en/la/safety
    Here we see “cadus”, “securitas”, “salvatio”, “salutem”, “sospitas”, “gratia”
    You may need to convert Greek translations into Western alphabet characters, such as
    https://www.lexilogos.com/keyboard/greek_modern_conversion.htm
    “asphalia”, “sigouria”, “perithalpsi”, “vios”, “zi”
  3. Find a 3-5 letter root word (in English, Finnish, or any other language) that you could imagine building a name on. Ideally one that hasn’t been well used already.
  4. Go to https://auctions.godaddy.com/ and search for all available .coms that have a maximum of eight or nine letters that include that root word. If a domain name is any good, someone will have found it already and will be trying to sell it.

Companies that appear to use a similar process to this include the shoe company Zappos, from the Spanish word “zapato”.

Keep looping. I’ve tried a lot of methods and this one has been the most fruitful. Any nice sounding, reasonably short, easy to spell .com will be taken. Auctions allow you to buy one of them for $1k-$3k. You can also search for more freely available domain names of course, but my experience that is you’ll be limited to 10-13 letter names.

Say the name to people. How does it sound being said? Do people ask you to spell it?

Don’t rush into naming. You can keep yourself under a working name for a while. You can even incorporate and contract under a working name and update contracts later on. Make sure the mission and direction of the company are tightened before committing to a name. The cost and burden of re-branding later on are worth taking some added caution up front. In my current startup, we operated for about a year under a broad name (such as “Orange”) that we liked but which was so commonplace, we knew we’d have to at least tweak it to get a trademark.

Do you need a .com? Perhaps not. There is a trend to naming companies “tryorange.com” or “orangeapp.com”. The only issue here is that the odds are low that you’ll get Orange as a trademark in your industry and if you start branding and launching without a trademark, you’re liable to a lawsuit later on that will force you to change your name. If you can get the “.com”, odds are reasonable you can also get the trademark.

Working with a copyright/trademark attorney to register your name is a good step. They will search copyright databases that you can also check: http://cocatalog.loc.gov/

. . .

All books and other resources referenced in this article

How to Develop the 2nd Product Your Startup Needs: A Sales Engine

This is Part 6/18 in the series “How to Build an Innovative New Product or Company” on the topic of building a sales and marketing machine to (a) find customers and (b) teach them how to buy your product

There are two reasons why startups fail:

  1. Too much focus on a technology and not enough on a buyer and user need
  2. Too much belief in the mantra, “if I build it, the world will beat a path to my door”

The “4 Steps to Develop a Strategy” codify a simple process that makes sure #1 should never be an issue. Codifying a failproof approach to #2 is trickier. How are you going to find your buyers? Do they know they have a problem? How are you going to educate them on it if not? Are they aware of your unique solution? How are you going to educate them on it if not?

In a startup, you’re either building or selling (or both)
In your first couple of years and while you are in your first 10-15 employees, everyone should be building your products and/or selling. Those two roles parallel the two reasons startups fail listed above. It’s also why your startup needs a second product—a sales engine.

Who are our potential customers? How do we categorize them? Do we want to prioritize some more than others?
Geoffrey Moore’s great book “Crossing the Chasm” helps us here. He shares how there are five types of people when it comes to technology adoption:

Innovators

  • They want to be the first to use a technology. They will forgive its flaws easily and evangelize the most innovative aspects. They like to tinker and play around with new things. They love Segways because of the technology; they’re not bothered that Segways can’t climb stairs and thus have limited use cases in the real world.
  • How do you spot them? They are interested in the analytics and the technology much more than the business case (though they may appreciate others in their organization will need to see the business case).
  • Want to alienate them? Tell them you already have a hundred customers using your product—they’ll lose interest and go look for something that is so cutting edge it hasn’t been so well discovered yet.

Visionaries

  • They have a business problem they need solving and are willing to try unproven, but promising, new approaches in order to make major breakthroughs. They are dreamers and want to be seen as leaders in their field by tackling the biggest problems that they and their peers face. Market leaders who are also Innovators are the most powerful first customers: be willing to invest as deeply as you need to sign them up—their reference-ability is what will allow you to sign Pragmatists up later on.
  • How do you spot them? They talk about a business problem they need solved and lament how there’s no solution out there yet to do it. They like that you have no competition because you so innovative.
  • Want to alienate them? Talk about the technology/features/functions, but not the business problem.

Pragmatists (on the other side of the chasm)

  • They want to be innovative but have lower tolerance for risk. They typically only talk to others in their industry; they are unlikely to value innovative ideas of someone that sits across industry boundaries.
  • How do you spot them? They will ask for case studies, proof points, and referenceable clients. Thus, they can’t be among your first few customers—early adopters have to fill that role. They like to hear “industry standard” in your pitch. They want to see that market leaders are already using your product.
  • Want to alienate them? Not being super buttoned up about what problem you solve, why others are using it today, why those clients love it, and leading them step-by-step through a well-known buying process. Telling them you are “state of the art”. Telling them you have no competitors; therefore, you should “invent” competitors if needed.

Conservatives and laggards

  • We’ll merge these two groups into one because the message is the same: startups should avoid them. Spot them and move on as quickly as possible. Don’t spend time trying to sell to them unless you have 10%+ market share; doing so will simply frustrate both of you.
  • How do you spot them? They will ask about your market share, how big you are, how stable you are, why they should use your product relative to doing nothing, how much support you offer, and they will continually question how simple the product and buying process really is.
  • Want to alienate them? Be a startup; be unable to provide clear answers to the above questions.

For the most part, startups want to find and focus their time on Innovators and Early Adopters. Those are the buyers who are looking for the type of innovation you provide and they will be great testers and feedback partners. The “Chasm” that most startups fall into and fail to cross, the thesis of Moore’s book, is the step where you have to rebuild your messaging, marketing, sales, and support/service functions as you cross out of serving pure innovators to serving those on the other side of the chasm, who are an entirely different audience and have entirely different buying needs.

When we productize our approach to finding customers, we first need to know which category any prospective customer fits into. We need to know where we are in our lifecycle and thus, we need marketing, sales, messaging, and outreach materials that provide the right message to the right person.

Should we go after a broad customer base (e.g. international customers or people who use our product for two or more different purposes)? Or should we be focused?

Moore offers advice here too. When you are getting your first few customers and your first few million $ in revenue, you need to be focused. You need to serve customers who will be a reference for future customers—and you should keep a pool of customers (and prospects) as tightly constrained as needed such that all of your customers could be meaningful references to any other customer. In other words, don’t go international if an international customer would neither value a reference from, nor would later be helpful in providing a reference to, a local customer. Don’t go after or serve customers that use your product to solve problem Y when such a customer would neither value a reference from, nor would later be helpful in providing a reference to, any other customer who uses your product to solve problem X.

How to find potential sales leads
Here are some ideas that have worked for me:

  • Establish an Advisory Council of Market Leaders and Visionaries. Get their feedback on the product and once they are excited enough about your direction, ask if they’d be willing to make introductions for you.
  • Build a “refer a friend” function into your product.
  • See what connections you have in LinkedIn to potential buyers; whom do you know who could make a customer connection? Offer a finder’s fee if introductions are hard to get.
  • Invest time in thought leadership. Publish and share insights that buyers will pay attention to. Become a leader in the value pool you’re targeting.
  • Used LinkedIn to find people who are in the right role and current organizations.
    • Further filter based on where they used to work. If you’re trying to find innovators early on, it may help to find functional leaders who have moved industries—they can be more respected by their current executive teams as sources of innovative new ideas.
    • Can you find potential buyers who moved into a new role in the past ninety days? They may be looking for opportunities to have an early impact (the downside is that they may also not know how to navigate the buying process in their new organization).
  • If you have to send emails to someone cold (such as from an email list of from LinkedIn connections), send a short note in plain text, sharing credibility (e.g. customer impact) and asking for a referral to the right person in the company.
  • If you happen to catch them by phone, ask “Did I catch you at a bad time?” “If you were me, how would you approach your organization?”
  • Offer to go at risk. If your product is $100 and, in the early days, you don’t have the prior customer proof points, offer to sell it for $50 if they’ll put $50 in escrow. Then agree on an outcome (or process) measurement such that the escrow goes back to them or to you at the end depending on if the product was successful. This is better than offering a trial period—typically, buyers and users aren’t committed in a trial period.
  • In my experience, it’s as much work to give a product away for free as it is to get someone to pay. And people tend to undervalue what they don’t pay for.

Aaron Ross, in his book “Predictable Revenue”, tells the story of how he led Salesforce.com’s initial $100 million of revenue growth. He offers a few words of advice about sales lead generation: that it shouldn’t be done by salespeople. Salespeople are among your most expensive resources; they should not be doing your most commoditized activity. Their Rolodexes will help in their first few weeks, but what you need is a sales lead generation process that creates leads for all salespeople and is not subject to their own abilities and capacity to do lead generation.

It can take a few months to get an inside sales-based lead generation engine going, but he recommends investing in one once it’s clear you have Product-Market Fit. Dedicate a full-time role to prospecting.

Prospecting should also help prune the list of qualified leads, not just create them. Are there warning flags that hint that a potential prospect is going to be a waste of time trying to follow up with? Ross gives example flags such as, “They just installed a ____ kind of system” or they are “Know-it-alls”.

Teach your customers how to buy from you

My experience has mostly been in B2B SaaS enterprise sales to health systems. Arguably the tools you need to be successful there may be overkill for B2C, non-enterprise sales, or sales in other industries.

One devilishly important lesson is the understanding that, simply because a prospective buyer likes and wants your product, that doesn’t mean they know how to get it. It’s up to you to find a sponsor and lead the sponsor through the steps. If you’ve sold your product to ten other customers, undoubtedly you know far more than any prospective buyer at the eleventh customer about how to run the process, how long it takes, the types of roles who need to be included, and what the pitfalls are likely to be. And, frankly, you probably have far more time and mindshare to give to the process of closing the deal than the busy executive on the other side.

What you need is a playbook that you and your sales team members (as your grow and hire them) should consistently use and rely on. The playbook reminds you where you are in each conversation and what the upcoming steps should be and it provides tools you can give to your counterpart to help them through the buying process.

. . .

All books and other resources referenced in this article

Does Your Product Only Need to Change the Behavior of a Few Users?

This is Part 5/18 in the series “How to Build an Innovative New Product or Company” on the topic of understanding what type of changes in behavior you need some or all of your users to make in order for your product to have the impact you expect and be successful

A few examples

“Change a Few Users a Lot”: Using e-learning to improve clinical outcomes in high-risk hospital-based events
In a prior company, we had a software-based platform to help doctors and nurses stay up to date with the latest standards of care in rare, high-risk events. Its goal was to improve their individual decision-making abilities, which then improved overall clinical outcomes. We sold our platform to hospitals under the premise that learning was needed by all clinicians on at least an annual basis. But how true was that?

Part of our e-learning was an assessment that was given to the clinicians before they took any learning. How they did on that assessment determined the learning they were then prescribed by the platform.

It’s true that the primary goal should be for everyone to improve a bit. Doing so creates a culture of vigilance. But a quick look at the assessment scores as they came in made it clear that it was a small group of low-scoring clinicians who created most of the patient safety risk.

So, while we publicly celebrated the incremental improvement everyone made on average through the e-learning, privately the discussions focused on what support and additional interventions were needed to shore up the small group who needed a lot more help.

This example is a case where 5% of the users caused 65% of the risk. You can also say 20% of the users are causing 80% of the risk, also known as the 80/20 rule. It’s an example of the Pareto curve, also known as the power-law distribution. It appears just about everywhere.

“Change a Few Users a Lot”: Inventing a technology to reduce car emissions
In another example, imagine you wanted to build a product or service that would reduce car emissions. You’d probably start by thinking about what you could add to every car on the road. But, as Malcolm Gladwell shares in his article “Million-dollar Murray”, 5% of the vehicles on the road contribute 55% of the pollution, according to a Denver study.

In this case, if you want to have an impact, you first need to figure out which cars these ones are (e.g. older cars or sports cars) and design a process around them. Sure, you could design something for every car out there. Perhaps doing so would lead to higher profits because you’d have a larger market to sell into. But if your goal is impact, you can’t ignore the power law.

“Change a Few Users a Lot”: Reducing the costs of the homeless
Another example in Gladwell’s article discussed the medical and policing costs incurred caring for the chronically homeless amounts to tens or hundreds of thousands per year, per person. But that’s just for those who are chronically homeless; they account for just 5% of the homeless population at any point in time. If you want to put a real dent in reducing homelessness and the cost society absorbs from the homeless, you have to start with this group.

“Change All Users a Little”: Connecting people via a social network, e.g. Facebook
Facebook is an example where, for the product to be successful, you need a large number of people to all make a small change in their behaviors (e.g. to login to Facebook and read/like/post something every day or so).

If 20% of your potential user base has to change their behaviors a lot to get 80% of the total impact …

Does your product need a component where you identify who the highest impact 5%, 10%, or 20% of users are? Is your product only going to be used by a small set of users?

If either of these are true, how does it change your market sizing? If this is a B2B product and the executive buyer realizes that only a few people on their staff will need to use your solution, will that affect your pricing?

From a pricing perspective, the ideal case is that everyone still has to use your product and you are able to claim value by the small behavior changes that result from that. Even if everyone uses your product, you would be well-advised to develop special features and interventions (on-platform and off-, such as via a client services team) specifically for the top 5%, 10%, or 20%.

How likely is it that this small set of users will change their behavior a lot? You may have to invest in hands-on client service support to get them there. There’s a general rule that, to have a product be successful, it cannot require users to change their core behaviors.

If everyone in your potential user base to change their behaviors a little bit to reach the total impact …

You need a pure product-focused solution. In the above case, the solution likely has to be a mix of products and hands-on client care support teams. Here, client care support teams won’t be as valuable as they simply can’t scale.

Such a product needs more investment to onboard users and train them on the new behaviors built into it. Such a product may find value in making social connections so that users can see the behaviors of other users as social proof.

. . .

All books and other resources referenced in this article

Finding New Area of Growth: Lessons from Disney and DC Comics

This is Part 4/18 in the series “How to Build an Innovative New Product or Company” on the topic of how to find high $ value pools, as discussed in the “4 Steps to Develop a Strategy”

If you have an existing user and are looking for future growth opportunities, then you have two vectors to consider:

Same users + New products
e.g. find new value pools/solution for the same buyers or users you currently serve. For example, selling a Batman movie to Batman comic readers.

New users + Same product
e.g. extend your product to serve more users within your customer base, or find new buyers for your product, such as in other industries

Note that I differentiated the two choices using users, not buyers. You could structure the different options between selling a new product to the same buyer, and vice-versa, but I find considering it in terms of the user most helpful here. Specifically, because the question here is whether you want to build a platform or not.

Your product is a platform when you have earned the right to all events relating to either (a) a particular set of users or (b) a particular business objective. Over time, you can then add more capabilities.

Google is an example of (a). It has buyers (advertisers) and users (you and me). It is a platform in that once it has us as users, it prefers to offer us more (e.g. email and maps) rather than find ways to provide its core product, search, for new users (e.g. building Google search into Wikipedia). Once Google has earned the right to be in our lives as a user, it has a built-in advantage when offering us new products, especially if they have similar components, a similar look-and-feel, and/or if they talk to one another.

More common is (b). Every company has an HR system that is the single source of truth for any workflow relating to corporate HR functions: who the employees are, where they live, when they were hired, who their manager is, and so on. CipherHealth declares themselves a “patient engagement” company and is a platform for any workflow related to that business objective, which includes surveying patients when they are in the hospital on how their care and stay could be improved as well as offering training and follow ups to them after they are sent home.

One question you might want to ask yourself is: where is the battle for users’ attention likely to be greatest over the next few years. By growing in one direction (e.g. by building a multi-product platform for that user), will you have an advantage and a head start over the competition?

There are many examples of either option. One that comes to mind is DC Comics. Long stuck with a specific type of user and a specific type of product (paper-based comics), they expanded both dimensions in the 1990s. With the movie “Batman”, for example, they were able to cross-sell their current user base to a new type of product—but, more importantly, they were able to dramatically increase the types of users they were able to reach. It should be added that DC Comics was influenced in this approach by Walt Disney, who articulated it very clearly in the 1940s and 1950s: develop and leverage the connections between Disney movies, music, merchandising, comics, and experiences (e.g. Disneyland).

Some questions to help decide which direction to go in:

  • Where can you get best cross-product sales leverage?
  • Are you being pushed by the market to be more than a point solution? In other words, are your customers saying it’s too much to use your product and another product at the same time?
  • Who are your likely acquirers? Position yourself to complement them but not compete; don’t partner with others who might take you in a competitive direction.
  • Who are the major players in your industry and what are they likely to go into? Where are platforms being built and along what dimensions? In healthcare, there’s a powerhouse company Epic which started as a well-constrained Electronic Medical Records company. As it grew, it built itself to be the one-stop-shop for anything that interfaces with physicians and nurses in a hospital—i.e. it’s a platform. While it does allow third party apps on its platform, it also builds its own capabilities onto it. It is so powerful as a platform that when Epic even hints it is going to develop a particular product, any company that has a similar existing product immediately gets pinched.
  • Where can you build a platform? Where can your data streams and buyer/user access allow you to easily add on additional use cases that buyers will pay more for? One reason to position yourself as a platform is that in an acquisition or fundraise, your valuation may get “credit” for potential revenues that you have not yet established if it appears all you need is to fire up a product team to build it out on top of what you have—and there would be an easily accessible customer base waiting for it.

You might think it unusual to be considering future growth opportunities as an early step in the process to build a single company and product. But I’ve found that type of thinking is at the heart of strategy.

A strategy that leads you to a single product is one thing; a strategy that spells out the direction for a multi-year, multi-product roadmap that slowly builds a powerhouse platform while no one is paying much attention is something quite a bit more powerful. A single product can be attacked by a larger competitor encroaching into your space in a few years; a platform, on the other hand, gives you the rights to land that you are playing on today and may choose to play on in the future. All of the startups I’ve been a part of have conversations about long-term growth vectors on a regular basis. This aspect of strategy is the one that is typically most heavily informed by knowing who the other major players are and what ground they’re staking.

It is also a great way to pave the foundation for a mission and vision statement for your company.

. . .

All books and other resources referenced in this article

An Executive’s Guide to Implementing AI and Machine Learning

Some lessons I’ve learned applying AI/Machine Learning to support business objectives

As a Chief Analytics Officer, I’ve had to bridge the gap between business needs and data scientists. How that gap is bridged is, in my experience, the difference between how well the value and promise of artificial intelligence (AI) and machine learning is realized. Here are a few things I’ve learned.

AI = machine learning (at least in 2019)

Machine learning is a path to get to AI. At least as of 2019, it is the only known viable path that I’m aware of. In the coming years, there may be other approaches. The two terms are not interchangeable, but for our purposes I will focus on machine learning.

Machine learning is a category of tools and approaches where a computer is given a large training set of data that includes an “answer key”. The machine then learns how to derive the answer key from combinations of the inputs. The model is then tested against a different testing data set to determine its accuracy.

Machine learning as a category can include basic statistical tools (e.g. linear regression) that fit this approach. It also includes neural networks, decision trees, and several other tools.

Is machine learning the right tool for the problem you’re trying to solve?

This one has tripped me up in the past.

For example, recently I had a data set with a lot of data collected from hospitals which had, for each employee, fifty measurements (for example, whether they showed up for work on time or whether they were consistently the only experienced person on their shift) and an indicator of whether they resigned in the weeks and months following. The question was: given this data set, could we create a model to predict employees who would resign before they did so, allowing hospitals to intervene early?

We spent months reviewing the data set and had used basic data visualization approaches to determine a set of rules. For example, employees who were just hired were twice as likely to resign than employees who had already worked at the hospital for ten years. Employees in certain clinical specialties had different average resignation rates as did employees in specific age ranges. Employees who were asked to work over sixty hours a week for two or more weeks per month were 50% more likely to resign.

Was this problem a good candidate for machine learning? Couldn’t we just assemble these rules and build a statistical model?

Ultimately, I believe it was a good candidate, but a similar problem might not have been. Caltech Professor Yaser Abu-Mostafa describes three qualities that problems need to have to be candidates for machine learning:

  1. There is enough data.
    • Machine learning becomes a better approach, relative to other options, the more data you have to train it. If you only have a few hundred rows of data, it may not work or may not work effectively.
  2. There is a relationship between the inputs you have and what you’re trying to predict.
    • In my example, our visual review of the data showed there were a lot of those relationships.
  3. The pattern cannot be described in plain English.
    • In my example, this one is less clear. Because we were able to describe many of the patterns in plain English, why couldn’t we just build a model on those patterns? In our case, there were both hidden patterns that we could not discern and relationships between variables that we could not discern or describe. For example, if someone is both working many hours and the most experienced member of their shifts, is the impact of both additive?

Starting here and making sure that the right tools are deployed might save you a lot of time. Even if the above are true for your data set, are they true enough to justify the investment in the approach?

From an ROI perspective, is the value of predicting an event that happens = to the cost of erroneously predicting one that does not?

Data analysts compare and optimize models based on accuracy. Accuracy is a tradeoff between predicting as many of the events that will happen versus incorrectly predicting as few of the events that will not. There are ways to adjust the model thresholds to see the tradeoff after the models are run, but interpreting such results requires a bit of mental gymnastics.

A better approach for business owners is to set up the models before running them in the terms of the business objective. In our example, the value of correctly predicting an employee who will resign might be worth thousands of dollars (i.e. the value of avoiding having to replace them with a new hire) whereas the cost of incorrectly predicting the resignation of an employee may be small. However, there are thresholds: if 10% of the workforce resigns every year, our understanding of our customers informs us that we can’t identify more than, say 20%, as high risks to intervene on. These criteria need to be translated from business requirements and vernacular (e.g. ROI) into model inputs (e.g. penalty matrices).

The important point is to determine the relative cost versus value of this tradeoff and to ensure the model is built to optimize on it.

One of the most famous machine learning implementations, the Netflix Prize, overlooked this I believe. The prize was given to the model that could most accurately predict movies that a viewer would like. Accuracy in their case was impartial to whether we gave a low score to a movie they might love or whether we gave a high score to a movie they might dislike. Is that the right business need though? If Netflix recommends five movies to me and I know I will strongly dislike two of them, as a user, I may well discount any of their recommendations in the future. But if I see five movies recommended for me and it doesn’t happen to include a movie I know I’ll love… that’s really not an issue for me at all.

Do you need stability in the model? In other words, if one small thing changes in the input, is it OK that the output swings wildly?

In my experience, the more accurate the model (at least by machine learning’s definition of accuracy), the more discontinuous the input-output relationship.

If you have a series of photos and you want a machine learning model to flag the ones that include a fire hydrant, you may well be OK with this level of instability because there’s no case where a user will change a pixel in a photo and expect that the output will remain similar.
In our case, though, some amount of stability is important. If we’re following an employee over time, we don’t want their predicted resignation risk to jump unpredictably every month when their age goes from 31.1 to 31.2 to 31.3.

Do you need to know which input fields are contributing to the predicted output? Or is it OK to have a complete black box?

Neural networks are black box models. You give it your inputs and it will return an output; that’s it. You get no insights into how it got the output or which inputs were most heavily weighted in the determination of the output. Neural networks tend to be more accurate than other models, but is the extra few percentage points of accuracy worth the complete lack of transparency to you from a business need?

Other models, such as linear regression and decision trees, will show you the combination of inputs that lead to the outputs. Data analysts may want to run multiple decision trees and take the average of all of their outputs. This may increase model accuracy but may reduce the ability to see how it is working.

I’ve found mixing models to be a way to get the best of both approaches. One model is a neural network model (with some dampening factor to maintain stability) that gives the turnover risk value. Another model is a linear regression that identifies the key drivers. We display the results of both to users. There’s some risk in that there’s no guarantee the models will match up (in other words, the neural network may identify a high-risk event that the other model does not identify drivers for) but there are ways to account for this.

In conclusion …

The above question and discussion sections were based on experiences I’ve had trying to bridge the gap between business objectives and analytics teams. Business objectives require a mix of both art than science. The science behind machine learning is increasingly commoditized: models are available for anyone to run and there is a wealth of resources available for getting to mathematically robust results. But the business you’re building is not a mathematical one. It’s a mix of user needs, marketing needs, and practicalities of integrating and deploying solutions. The art of deploying machine learning solutions is particularly interesting to me and hopefully these give some points for consideration.

. . .

All books and other resources referenced in this article

Some Strategy Mistakes I’ve Made

This is Part 3/18 in the series “How to Build an Innovative New Product or Company” on the topic of over-investing in the “4 Steps to Develop a Strategy”

I was introduced to a medical device company recently. They have created a product that allows you to do blood analysis faster and simpler than any other product out there. The product is a piece of paper that has six testing areas on it, each of which can test for something different. It’s relatively cheap to make, can be easily carried “in the field”, and is disposable. They’re starting to think about how to sell and market it. Sounds like a winner, right? Perhaps, but what’s incredible to me is how commonplace this type of company is. They have a technology that, to the founders, is self-evident as a game changer. They just haven’t figured out yet whose game it’s going to change.

A company like this has a product and a unique technical insight. But they don’t yet have a strategy. Who is their customer? What is their need? Why can’t two kids in a dorm room build this product?

There’s nothing wrong with starting a company this way.

A great strategy and path to market may well exist. But things go don’t go well when you start with a technical insight and proceed forward, without pausing to do the hard work of developing a strategy.

For this company, they could now begin to enumerate all of the potential customers and their needs. For example, they could go and talk to hospital labs. What are their main needs? How big of an opportunity can be unlocked by the product? The potential buyers might say, “well, your product is as good as our $50,000 analysis equipment, but we’ve already bought it so a cheaper option won’t save us anything until we need to replace it”.

They could talk to EMTs and mobile labs. How much faster or cheaper would their test be than what potential customers currently use? How much value do they unlock by providing faster or cheaper results? Do EMTs wait to do any blood testing until the patient is stabilized in an ambulance and have no need for faster results?

In short, the story I see so often is a group of technically-oriented founders pushing headlong into a product without knowing who is going to be the first customer segment to use it, why they’re going to use it, and how badly that customer needs the solution.

Many product ideas have a customer identified… but not a fact-based understanding of their most urgent pain points and the under-leveraged root causes

Mistakes I’ve made … story #1
A couple of years ago, I had a need in my job. As a strategy leader in a prior company, I was looking for companies that were potential competitors or acquisition targets for us. I had found a few, largely in an ad hoc way, as colleagues pointed them out to us. How could I find others?

I inferred that M&A leaders in large companies and venture capital firms must have the same challenges. And I saw an opportunity to build a unique analytics solution. I built a tool that web-scraped the websites of over a million corporate websites and compared the terms that they used to speak about themselves. The site was called LogoSleuth. I’ve since taken it down, but at the time, you could go to it, search for a company URL (such as your own) and it would report back to you the other corporate websites that used similar wording about themselves as the one you were searching for.

It was a nice piece of technology. And I had a type of customer who could be a buyer: those who were in a similar role as I was. But I hadn’t spoken to them. I hadn’t gone through the hard work of determining who they were and where they worked. I hadn’t talked to them about their largest, table-pounding challenges. And I hadn’t then investigated the causes and current solutions to those challenges. If I had, I would have discovered that the need I had was a bit unique; that for those who faced it regularly, there were other tools. And overall, it wasn’t one of their top three needs.

Would it have been better to find out what those top three needs were and try to build a unique solution to them, even if the solution was a bit less perfect than the “solution” I had created? Yes … there’s no question. But sometimes you need to learn your own lessons yourself.

Mistakes I’ve made … story #2
About ten years ago, I led product development for an analytics startup. We had great technology that could identify areas of cost reduction opportunity and revenue growth opportunity for hospitals. The products were quite good—but we had about a dozen of these products. Our sales team had a tough time selling the individual components (they were too niche to get executives’ attention) and so our sales team embarked on a process of bundling products to create a complete solution that was defined entirely in terms of a buyer’s need. It was successful but painful. There’s nothing wrong with going through the “4 Steps to Develop a Strategy” backwards (start with a technology and areas of competitive advantage and look for buyers who are best served by the product) but doing so after we had launched and scaled up a sales team was a misstep.

Sure, strategy should be an organic, evolving construct inside of a company, but having a strategy in advance that meets the criteria of the four steps (even if it’s not quite right) really helps with the journey later on.

In summary…

I included this (the importance of having a strategy based on customer need) as step #3 in a series on the process of developing a company or product for all of the above reasons.

If you don’t have a strategy yet, you might find the articles I’ve written on creating one (referenced below) helpful.

. . .

All books and other resources referenced in this article

Find People with No Teeth to Inform Your Product Design Process

This is Part 2/18 in the series “How to Build an Innovative New Product or Company” on the topic of how to conduct buyer and user interviews to help shape product definition, innovation, and development

We don’t give lip service to consumer understanding. We dig deep. We immerse ourselves in people’s day-to-day lives. We work hard to find the tensions that we can resolve. From those tensions come insights that lead to big ideas.

— former P&G CEO Bob McDonald

There are two diametrically opposing goals of user interviews

You have to know in advance what your goal in an interview is or you risk confusing the process. You can use interviews to either …

  1. grow the potential solution/ideation space by more deeply understanding users’ needs or
  2. narrow down the space by zeroing-in on a specific direction

When I was strategy consultant, I supported a colleague who was doing a few expert interviews to inform the direction of a strategy he was developing for a client. As consultants, we did lots of interviews. We’d rely heavily on talking to experts. Not because they could do the hard work of strategy creation for us, but rather because they are such a direct way of validating or disproving hypotheses that informed the direction we were taking.

In this particular instance, the consultant did three expert interviews. Two of them confirmed his hypothesis and the third was challenging: it raised nuances about why it might not be true. The consultant colleague of mine ended the phone call and said “well that guy was a waste of time.” And, I suppose, there was nothing inherently wrong with him saying that. He was on a tight timeline and had all the evidence he needed that he was directionally correct. He didn’t have the time to get into nuances. Sometimes the key to strategic thinking is to rise above details and make bold moves, knowing that when any innovative idea is let loose into the real world, it will encounter some degree of friction. His goal at that time was to gather enough evidence to defend zeroing-in on a specific direction.

What matters is that you know your goal. Having the time to deeply understand your buyers and users is tremendously valuable. In that scenario, your goal is not to get simple answers; it is to develop real relationships and to get a complex appreciation of users’ lives and motivations. You want to grow your set of potential options and challenges (e.g. through open-ended, empathetic user interviews). Only later should you prune them back (e.g. through more decision-oriented expert interviews) when you need to align on a specific and well-informed path.

IDEO has written several great books about this empathetic, deep-immersion approach to innovation. In “Creative Confidence”, IDEO partners David and Tom Kelley share the classic story of how the project lead of GE’s MRI machines, basking in the glow of a new products’ technological perfection, immediately saw a young patient crying because he was scared to go near it. He took a journey that led him to redesign GE’s pediatric MRIs visually and to create storylines around them, such as the patient sailing inside a pirate ship. It’s a perfect case study where the product was initially designed for the buyer and the technician with no insight into (or empathy for) the user.

Conduct customer interviews to develop empathy for your users and to grow new ideas

First: whom do you want to talk to? Who are your users? The answer may be clear—or you may need to consider what mix of age, gender, career level, geographic location, or other factors you need.

Come to the interview with an interview guide: a set of questions that you step through and document answers in. Questions are ideally open-ended when you’re in exploratory mode, not yes-and-no. You can allow variation from the guide as needed, but it will keep you focused and make sure that results from multiple interviews can be aggregated into coherent insights. Your interview guide should evolve as interviews progress: certain questions will be answered easily and new areas to explore will be discovered. Have an interviewer and a separate notetaker; write down what your interviewee says, with real quotes, not just what you interpreted their answer to mean.

How to find interviewees
I’ve used GLG and similar firms in the past to help me find and schedule experts to interview. If you’re looking for CxOs in Fortune 1000 companies and are willing and able to pay for such access, they are a great option. For a startup, you may need to look at other avenues.

LinkedIn has been useful. Most business leaders are on it and it’s a great way to find Director- and VP-level employees who are easier to reach and more willing to talk (and sometimes better informed). I usually offer an Amazon gift card in exchange for a one-hour expert interview. LinkedIn has a premium plan that lets you send InMails on their platform. Yield on such cold outreach is low though: a 5% or 10% response rate.

Tom Chi, formerly of Google X, shares a great insight in the video linked below: if you’re building a B2C product, go to the mall twice a week to get in front of potential customers. They can be a receptive audience because some people go to the mall when they’re bored so they are interested in seeing what you’re up to. And also, perhaps most important, the mall does the demographic segmentation for you: there are stores that focus on different groups such as skateboarding teenagers. Find the store that fits your user demographic and stand near it. Then have a couple of engineers sitting in the food court who are able to update the prototype in real time as you get feedback from users. Of course, this works if you have a relatively simple B2C product and are looking for initial feedback. It doesn’t apply as well to the challenges that I have found most interesting—such as deep usage questions (which require days of usage to fully understand), usage that requires being part of a users’ daily workflow, or B2B users (whom you’re not likely to be able to find in a mall, at least not in the right mindset).

You can’t ask users to evaluate or rate behaviors that they don’t do today
Find the leading users who have already found a way to accomplish the behavior your product is trying to enable and interview them. Why is it hard? How can we make it easier? For other users, ask them “Do you do this today?” “Do you try to do it today but have trouble? If so, where do you get stuck and why?”

Be aware of the “consolation prize” interview feedback
If you’re sharing an idea with a potential buyer/user and they don’t like it, don’t follow up with “Well about if we did this?” or “What about if we changed it this way?”. They’ll likely say, “That would be better!” because they don’t want to disappoint you. Ask them to disappoint you. Start new ideas carefully with a clean slate and ask for objective ratings. If you take the consolation prize answer, you’ll end up with a bad product and both of you will be disappointed later.

Be aware the pitfalls of the “Pepsi Challenge” in interview feedback
In the 1980s, Coke and Pepsi ramped up their marketing to win customers away from each other. Pepsi set up tasting tables where Coke drinkers would drink a small sip of Coke and Pepsi and apparently say “I like Pepsi better”. This was because Pepsi is sweeter. So, in small doses, they may prefer Pepsi. But ask them to drink an entire can every day and most likely their feedback would be “Pepsi is too sweet”. I’ve found the same thing when looking over UI designs. In small doses, people like bold and colorful. But when they have a product fully built out and are using it every day, they want less attention-grabbing graphics, less text, and simpler iconography.

IDEO puts special emphasis on “extreme users” (i.e. power users)

Who are the potential users of your product who are early adopters, tinkerers, and innovators? Who has already cobbled together a solution to what you’re building by themselves, using Excel and other available off-the-shelf tools? Who are the users for whom your value proposition is to simply help them save time and do more efficiently what they are already doing (as opposed to helping them do something they hadn’t even considered doing)?

Those are your “extreme users”. By observing and engaging with them, you will be able to fast-forward through a lot of product design decisions. Other “extreme users” are those who have specific needs that make existing products a poor fit—such as tall people driving small cars or elderly people with arthritis trying to open medication bottles.

IDEO, for example, interviewed people without teeth during the ideation for a toothbrush that was later marketed under the slogan “unless you have a flip-top head…” A senior partner at the strategy consulting firm I worked at a few years ago shared this story with me as inspiration for a series of interviews I was about to embark on in the healthcare analytics space. I went back to him later, “I just spent weeks with these toothless people down by the bus station and really got nothing usable from it…”

How to manage a 1-1 user feedback interview

Remote research (by phone or web conference) has several positives: people are easier to recruit, you can connect with people all over the world, it’s easier to have other members of your team join or listen in and it’s easier to keep your interviewees in their natural environment (e.g. their office) where they may be using the product you’re building.

Remember to ask them to be at their computer in a quiet room when you schedule the interview—otherwise some people will schedule you during their commute.

At the start of the interview, check that they still have an hour to spend with you; if not, ask to reschedule. You’re paying them. Don’t accept an abbreviated or distracted time slot. Ask them if it’s okay to record the session, which web conference tools like Zoom make it easy to do. Remind them not to share any confidential information about their employer during the interview.

Jake Knapp’s book “Sprint: How to Solve Big Problems and Test New Ideas in Just Five Days” has some great guidance from him and his colleagues who run the User Experience Research Group and Google Ventures. This is a topic they know something about.

Sitting side-by-side with a user as they are testing out a new user interface can be beneficial, but that’s only one type of interview. My colleagues and I do lots of interviews all the way through the product development process—from understanding needs to asking how a basic problem solution might fit into their day to seeing how they interact and use a prototype. Many of them can be done as well over a web conference—which will allow you to talk to a greater range of potential users more efficiently.

Jake explains five stages of an interview:

  1. Warm intro. Smile. Be positive and enthusiastic. Put the interviewee at ease. Tell them how you’re looking forward to getting their feedback and reactions. Let them know you’re still working on the product and so you know it’s not going to be perfect. Ask if it’s okay to record the interview for your internal notetaking later on.
  2. Ask a couple of open-ended questions to better understand their context. Jake recommends starting with small talk (“how long have you been doing this role?”) to zeroing in on a topic related to your product (“do you use any tools or software systems on a daily basis in your job?”). Open-ended questions are ones that can go in many potential directions.
  3. Bring out the idea, prototype, or product. Let them know “I didn’t design this, so you can’t flatter or offend me—please share honest feedback.” In a startup, you may well have been doing all the design and interviewing but any opportunity to put some distance between yourself and the product is helpful to keep the interviewee from coloring their reactions.
  4. Ask them to try to accomplish something with the product. Yes, you can tell them step-by-step what to click on, but what would you learn from that? Your product needs to be able to be useable with no training. You can ask “if you were seeing this for the first time, how would you decide if you wanted to spend more time with it?”, “What would you want to look for first? Why?”, and “What are you thinking when you see this?”
  5. Debrief. Away from the product again, “What did you like? Dislike?” and “if you could change or add something, what would it be?” Ask open-ended questions throughout, as opposed to, “would you likely use this on a daily basis or a weekly basis?”, which assumes one of those two options is viable for them. If you feel the need to ask a yes/no question (e.g. “did that graph look wrong to you?”), try an open one instead (e.g. “I see you’re pausing on the graph…”). Or show them two mockups and ask, “do you prefer one over the other? If so, why?”

Remember: the interview is about them, not you

This is a trap I’ve fallen into in the past. I’ll say things like “You’d never use this feature, right? Because …” or simply, “I agree”. Your opinions don’t matter. You’re there to learn and document. They’re the expert. If you want to hear your own opinions, set aside another time to talk to yourself. Saying less and reinforcing your view that they are the only expert in the room increases the invitation for them to speak and share.

Your job is to be the advocate for your interviewee, not for you or your product.

You can have two people on your side join the interview, but you should have one person be designated as the lead interviewer and another as the notetaker. If the notetaker wants to ask a question, they should hand it off to the lead interviewer who should time it and frame it in the right way.

After asking an open-ended question, allow some silence after their answer. Silence is an invitation for them to speak. If you’re always running up against them trying to speak yourself, you are reducing the power of that invitation.

. . .

All books and other resources referenced in this article

Fill Your Founding Team With “MacGyvers”

This is Part 1/18 in the series “How to Build an Innovative New Product or Company” on the topic of building a founding team

Some of the most important lessons I’ve learned were from the founder of a company that I was part of a few years ago. A remarkable man, the company he had founded about twenty years earlier had gone through several major iterations.

In the beginning, he was a medical illustrator supporting lawyers in medical malpractice court cases. He told me about his early days in the courtroom. He had no clients so he would put on a suit and hang around the courthouse with his typical optimistic vim. He got to know all of the medical lawyers by doing that and when he bumped into them, he’d mentioned he was “supporting a trial down the hall but give me a call if I can support you some time.” There was no other case down the hall but that’s what an entrepreneur sometimes has to do. You have to live in the world that you’re imagining and will it into existence.

Some people call it “fake it until you make it.” Perhaps—but one lesson I learned was that there was nothing faked about his approach. In his mind, he belonged to be in the courthouse and when he said he had a client down the hall, he was talking about a future client that he knew he would have. The minor dimension of time was irrelevant.

By the time I knew him, the company had evolved into a medical e-learning platform. When I first joined, my desk was right across from his office and his door was always open. If the sound of his phone calls didn’t remind me of his presence, every few hours, the sound of twenty-five-pound curling weights being dropped onto the floor did.

My view of salespeople at that time was that they should be helpful to the buyer but not overly burdensome to them. That was not our founder’s point of view. He believed so strongly that patients’ lives were at risk if a hospital didn’t use our e-learning that he saw it as his responsibility to will obstacles out of the way.

There was one poor buyer at a large health system who had begun the process of buying our software but could not have possibly moved fast enough. I’d watch him every day. He’d call and leave an urgent message on her voicemail. He’d text her. He’d call again. He’d send an email. He repeated this every day. Months later, she told us later how bothered she had been by it and joked about a restraining order. He was as unnecessarily repetitive as “24 Hours to Midnight” is as the title of a spy film. But he believed so strongly in his moral obligation that none of that affected him—and customers usually saw that too. It didn’t feel like salesmanship.

Twenty years into his company and he was still hustling for customers like a Moroccan street vendor. Very few people ever called him back but that didn’t matter. He preached his version of the Gospel every day for so long over time that many people rallied behind him—employees and customers.

Whom you work with makes all the difference. It really matters who the first few employees are for a new venture.

Startups typically have two to three founders, though my current venture has four. Most of the founders I know have over ten years’ work experience and have been part of other startups in the past. There is an image of the college dropout founding companies but many are mid-career.

I think you need three things in your founding team. You need:

1) The smallest set of initial team members who both complement each other’s skill sets and cover the fundamentals needed to launch the business.
You need at least one person who deeply knows the buyers and the users and has warm connections to them (you’ll need them to support sales and product feedback sessions). At least one person who is an expert in the major technologies and/or trends you are applying to solve a major topic for the buyers/users. Potentially someone who has a lot of strong connections with investors who can help make fundraising easy. This person need not be a full-time employee; this is a great spot for a Board Director.

At least one person needs to have deep experience as a product manager and/or building products in an agile, customer-centered way.

2) Founders who are scrappy and pragmatic. Founders who …
You need founders who hustle to get things done. Founders who, regardless of their role in the company, will go off on their own and find a customer in the early days by knocking on doors in their spare time.

Founders who are MacGyvers, to reference the popular 1980s TV show. Founders who chip away at a problem day after day, who can take one obstacle at a time and not be overwhelmed by the sheer expanse of problems yet to be solved. USAF Col. John Boyd said “A winner is someone that can build snowmobiles…” His observation was that snowmobiles are a collection of existing parts from other vehicles (sleds, tractor treads, a motor) and so someone needed to have the creativity to see the parts and imagine them as a new product.

Angela Duckworth’s book “Grit” attempts to capture similar ideas. She writes that high grit people make things happen; they are self-starters; they chip away at a problem for a long time; they don’t get discouraged by setbacks.

In contrast, in 1940, Albert E.N. Gray wrote in “The Common Denominator of Success”:

Like most of us, I had been brought up on the popular belief that the secret of success is hard work, but I had seen so many men work hard without succeeding and so many men succeed without working hard that I had become convinced that hard work was not the real secret … The common denominator of success — the secret of success of every man who has ever been successful — lies in the fact that he formed the habit of doing things that failures don’t like to do … [For example, salespeople] don’t like to call on people who don’t want to see us …

Of course, hard work and perseverance (i.e. grit) matter. But I believe the types of work people are willing to do matters much more.
Build your founding team full of people who roll up their sleeves and do the hard work that most people don’t want to do. Those people do that hard, sometimes unhappy work because what drives their day-to-day behaviors is a focus on long-term results, not the satisfaction of achieving near-term gains.

3) Founders who have worked together before
I’ve almost always worked on my startups with the same small group of people. Some come and go over time, but the team is consistent enough that when we’re building something new, communication can be implicit.

For fundraising, showing a core team that has worked together is a major bonus. If the team doesn’t know each other, there’s a more significant risk that they’ll fall into one of the most common pitfalls of all startups: self-implosion. Investors often invest more in the team than the product. The reason is that there’s two options for any product: it’s either a brand-new idea that no one has seen before or there are several similar competitors out there. For the former, investors will never get enough proof that something unproven is actually going to work so they can only bet on the team that will figure it out. For the latter, they’ll bet on the team that’s got the best chance of winning; the team that has the most experience building companies and the most expertise of the customer and the market. Either way, it’s the team.

Further, until you have worked alongside someone else, how can you really know they’d not only be a good hire, but also a worthy partner and co-founder? Some of us may have a missionary’s benevolence for seeing the potential in others, but that needs to be curtailed when assembling a founding team, it’s one moment that’s too risky time to gamble on.

. . .

All books and other resources referenced in this article

Strategy and Agility: Lessons from Toyota, Honda, Yamaha, and Tesla

A great example of strategy tightly coupled with agility/fast-product-feedback-loops: Toyota and the Toyota Production System

The aim of the Toyota Production System (“TPS”) is “making the vehicles ordered by customers in the quickest and most efficient way, in order to deliver the vehicles as quickly as possible” (emphasis is mine).

In other words, the TPS is the mechanism that Toyota uses to build a lean / agile / high mobility mentality into its production lifecycle from order to delivery. The TPS supports Toyota’s overall strategy of building the most reliable and lowest total-cost-of-ownership cars in the world.

All workers are obligated to improve any process they can to improve speed and reduce waste in the production line. They’re not just permitted to improve, but obligated… in other words, they’re not allowed not to speak up to their manager if they see an opportunity for improvement.

Why is speed so important that TPS’s objective mentions it twice? Faster design, ordering, and delivery of cars means Toyota can innovate faster, respond faster to customer needs, and have new ideas in the market faster than its competitors. Faster throughput also means less inventory and thus less resistance by dealers to push new models into the pipeline.

Workers need not run ideas up a chain of command, wait for a response from a committee that meets monthly, and then implement. They can decide on the frontline, in real time, when they are in line with the overall strategy and when speed can be used in such a situation to improve responsiveness to customers.

The Honda-Yamaha War

In Chet Richards’ book, “Certain to Win”, he shared the story of the Honda-Yamaha War, known as the “H-Y War”.

It started in 1981. If you were a young kid in Japan in the early 1980s, you likely wanted a scooter or a motorbike. Those were the dream products that reflected Japan’s newly-emerged engineering leadership, freedom, and economic growth. Yamaha opened a new factory with the expressed intent to become the world’s leading motorcycle manufacturer—a direct shot at Honda, the reigning leader.

How would a typical competitor, in Honda’s role, respond? With more marketing? By opening an even bigger factory? By cutting prices? Not in this case.

Honda responded as a lean startup should—by cycling as fast as possible through new models, learning from each, listening to customers, and using speed and agility as its principle strengths. Their new models evolved so fast that they influenced what the public wanted.

In 1981, Honda had sixty models. Over the next eighteen months, they introduced 113 new ones. Each model was replacing a prior one; each one was more attuned to their customers. While Yamaha’s new factory could produce a large number of models, their ability to keep up with the pace of change was limited. In the same time period, they introduced only thirty-seven new models.

Yamaha’s models quickly looked ancient compared to Honda’s.

How was Honda able to do it? They acted like a lean startup. They operated with a fast-product-feedback-loop so fast that Yamaha wasn’t even able to register the changes and updates they were making. Yamaha was so outmaneuvered they were left confused and disoriented to the point where they surrendered and announced, “We want to end the H-Y Way. It was our fault.”

Build fast-product-feedback-loops into your product outline (sometimes called “learning loops”) to create a flywheel of continuous product improvement

You’re looking to build feedback loops into your products and customer experience so that every user helps you get smarter over time.

Seeing what users click on (and what they don’t) are obvious UI data points, as are user feedback and rating scores.

Particularly interesting, though, is when users can contribute small bits of content that become part of everyone else’s experience; those products become especially valuable. For example, with the GPS-enabled app Waze, users can press a button while driving to share the location of a speed trap or a broken-down car. Waze then uses this information to alert other drivers. A click by one driver can create a valuable improvement in the user experiences of thousands of other users.

Google’s Nest thermostat claims to do something similar: when a few users adjust their thermostat, it uses that data feedback to determine how to improve its algorithm for managing all of its users’ temperature.

These are examples of a positive-feedback-loop: the more a user uses a feature, the more value is created in the system. Contrast that to most physical goods sales: the more a customer handles a book in a bookstore, the less new the book looks. And more customers reading a book doesn’t help make the book better for future readers.

Flywheels create network effects, which, once you have a critical mass of users become their own source of competitive advantage.

Tesla outdoes them all

Tesla cars send driving and user interaction data back to their engineers on a daily basis. The engineers review the data and update the operating system. The updates are then pushed out to all of the cars nightly. Tesla cars upgrade like a web-based platform does… constantly, with no work required by the user, and for no additional cost.

The changes Tesla is making to the UI on the touchscreen dashboard would require a complete rebuild of the dashboard for almost all other car manufacturers.

Compare that to the standard model of releasing a new version of a car once a year for those who buy the new version; those who don’t buy a new car every year get no updates. And even then, the annual updates are minor, largely uninformed by any actual data of driving activity of the prior year’s version.

In short, Tesla can update its touchscreen dashboard nightly while other car manufacturers require you to buy a completely new car.

When you can operate in a user-feedback-informed design and production cycle so fast that your customers and competitors can’t even register the change before you’re well onto your next update, you win.

Col. Boyd called this “getting inside the other person’s decision cycle”.

. . .

All books and other resources referenced in this article

Startups Should Do “Value Pool Sizing”, Not Market Sizing

Value pool sizing, as a mathematical exercise, can be simple. What’s the $ size of a pain that a buyer has? How much can we alleviate it? That’s it.

Your revenue is the percent of the value pool you can capture in price, while the rest goes back to the buyer as the value they get for using your product.

Note that a value pool is focused on a specific customer pain point and area of need. The need may not yet be served by any solution.

Market sizing is an exercise to size up … well, a market. A market by definition already has one or more major incumbents that define it. A market is then defined by a type of product and not a specific need.

How to do Value Pool Sizing

For example, let’s look at customer turnover for commuter rails. From our interviews, imagine we heard 30% turnover in rider base over the course of a year is normal. They measure the cost of customer turnover as equal to the cost to acquire a replacement—about $500.

Thus, the size of the pain = a ridership base of 100k people * $500 * 30% = $15M / year.

Our targeted impact is a percentage of that size of pain. If we can help them reduce their turnover rate by a tenth (i.e. 30% -> 27%) it would be worth $1M (2% * 100k * $500). That’s the value pool for this one customer.

Would the $1M savings per year get the commuter rail’s attention? You need to ask the buyers. And only through interviews will you learn, for example, that a Director of Operations role is solely accountable for managing all rider turnover.

Customers often look for a 5:1—10:1 ROI on their investments. Thus, the price point you’d target for this value pool would be about $150k/year. They spend that, and with a 7x ROI, they will see $1M/year in savings on an overall problem that costs them $15M/year. Maybe over time we can improve our product to take on a bigger chunk of the $15M.

Why I prefer value pools over market sizes

The exercise above is focused on a buyer, a buyer’s need, and how we can address that need. Market sizing, by contrast, is more about how much money customer are spending today on a product category. That’s just less relevant for a startup looking to introduce something new.

Value pools and market sizing can both be useful; they measure different things for different reasons.

If you choose to go the market sizing route, some thoughts follow.

TAM = Total Addressable Market

Let’s start with definitions:

TAM for Product X = the Total Addressable Market.

TAM is the total amount of revenue your product could make every year if you served all possible customers and they all bought a complete solution from you (Product X) including only product components you expect to have over the next one to three years if you made no major unforeseen investments or changes in product direction.

For example, if you’re selling sandals over the internet, you might size “Internet-based sandals”. But you might also size “Internet-based shoe sales” or “Internet-based beach wear”. This is based on an understanding that internet-based sales are core to your business model and it would take major unforeseen investments to change that (e.g. to open a brick and mortar store). But you could well expand into broader product category areas without unusual new investment. Which direction—all shoes sales or beach wear—should you size? Possibly both. The TAMs can then help you decide where to anchor your strategy.

When you calculate a TAM, questions may come up:

  • If I’m inventing a new market, how do I size it? (Answer: use a value pool approach; see above.)
  • What’s the right level of detail when calculating TAM for a product? E.g. do you calculate for food, candy, candy bars, coconut candy bars, or Mounds?
  • What’s the right level of detail when calculating TAM for a customer segment? E.g. do you calculate for direct customer sales, grocery store sales, direct business sales, or mix them all?
  • What’s the difference in size between TAM and SAM (the reachable, or Serviceable Addressable Market)? In other words, what part of the total market is serviceable by you (i.e. SAM) versus not (i.e. in SAM but not TAM) and why?
  • What if our product will grow over the next few years with new value-creating and price-increasing features—should we size based on today’s available market or the future’s?
  • What if we have customers who pay for an enhanced product—should we size assuming everyone might buy it, or size based on some mix of enhanced versus basic product options?
  • If we have an established product and an established price, should we use our price for the market size, our competitors’, or some mix?

There is no official answer to many of these questions. If you’re doing market sizing, as long as you align on an approach and use it consistently, that’s what’s most important. I guess I could offer more answers to these questions … but my point is, why bother so much with TAM anyway?

The nice thing about value pool sizing is that it sidesteps these and focuses instead on what really matters: how much are you able to help a customer.

. . .

All books and other resources referenced in this article