In early 2012 Sebastien Thrun, founder of online learning startup Udacity, claimed that in 50 years there would be only 10 institutions delivering higher education in the world. Thrun believed that Udacity could be one of them. Yet almost two years later in late 2013, Thrun himself no longer believed the hype. In Thrun’s own words from a 2013 Fast Company article,
“I’d aspired to give people a profound education–to teach them something substantial. But the data was at odds with this idea.”
The data at odds with his vision was that less than 10% of students completed courses on Udacity. To a man who wanted to educate the world, this data point was anything but evidence that Udacity was doing that.
Then, an epiphany. Thrun and his team realized that what people actually cared about was employment, rather than a ‘profound education’.
Again, in the words of Thrun, “At the end of the day, the true value proposition of education is employment. Why not give industry a voice?”
Armed with this insight, Thrun changed Udacity’s course to become a platform that helps users find jobs and helps employers train and recruit skilled candidates.
These learnings are clearly displayed on Udacity’s home page today:
This ‘pivot’ has driven Udacity to success - Udacity reportedly is doing over $100M in revenue.
It’s a great story, but did Thrun need to spend almost two years building a product that few valued? Did it have to take nearly two years to figure out that key partners in Udacity’s ultimate success would be companies looking to train and hire? Did Udacity need to burn through millions to discover the insight that had led to success?
I don’t think so. I’m confident that there is a way for entrepreneurs to avoid this circuitous path to startup success. Let me explain…
Developing a hypothetically robust business model around a hypothetically compelling value proposition is easy. Developing an actually robust business model around an actually compelling value proposition is hard. Given that the gap between the two can have significant consequences, I’m interested in how we can bridge the gap between the hypothetical and the actual.
Enter Customer Development
A well-known concept in the startup world is Customer Development. It’s a ‘lean’ method to develop a Business Model Canvas. A Business Model Canvas is a tool for developing and documenting your business model. The TL;DR on Customer Development is that you can improve the process of startup building by quickly iterating through untested hypotheses via talking to potential customers and prototype testing. To do this, you must “get out of the building” where the learnings are; the insights are not in your head or your office. You then document and structure your learnings with the Business Model Canvas.
I’m quite familiar with Customer Development as Allocate (a startup I co-founded in 2016 and sold in 2019) began as a class project in the Lean Launchpad at Stanford. That class, taught by Steve Blank, takes all teams through the Business Model Canvas via the Customer Development methodology.
We started the class by writing our business model hypotheses in the Business Model Canvas document. You’ll notice that the “value proposition” section is central to the Business Model Canvas. This is by design - a startup and/or product must have some value proposition that hooks customers. The other sections of the business model follow from that value proposition.
Then, over 10 weeks, we spoke to over 100 people about our different hypotheses in each section. We quickly learned that many of our initial hypotheses were incorrect, so we updated our priors and continued the learning. After 10 weeks we thought we had everything figured out - we only needed to raise money and build the product. Startup success was imminent.
How wrong we were. Customer Development is an incredibly useful framework that helped us go far, but it didn’t help us go all the way.
Here’s what went wrong - Customer Development relies on humans, and humans are not always reliable sources of information. Asking the right questions and getting the truth from your interviewees is challenging. Most people don’t want to tell you that your idea is bad and will typically ask for a faster horse rather than a car (we’re largely unable to describe what we want/need). Because Customer Development relies on interviewing humans to inform the business model, we begin to see how that business model can be off-base.
This is the root of the problem; it’s hard to bridge the gap between the hypothetically robust and the actually robust if one’s map, the business model, is built on inaccurate assumptions. Building a new business based on misinformed assumptions is like traveling to Paris, Texas when you meant to go to Paris, France. It seems like the right destination, and you feel like you’re making progress in the right direction (from California), but the route is different, and once you get there, well, they’re quite different places. (Though both have an Eiffel Tower - one adorned with a big red hat.)
Doing my best to be as clear as possible: Customer Development may yield a business model that, while hypothetically robust, may not be actually robust. And this can cause problems. In particular, starting a new venture or creating a new product based on inaccurate foundations can cause three issues that entrepreneurs would be wise to avoid.
One may build a bad product.
One may orient functions of the business around the wrong activities.
One may operate with less capital efficiency.
The way to avoid the problems listed above is to get the business model right, as quickly as possible, by improving Customer Development.
Thus, the main idea is that if one can, throughout the process of Customer Development, understand why customers care or don’t care about a product’s value proposition, then one can build an actually robust business supported by an actually compelling value proposition, and solve for the problems listed above.
Understanding Why
When I was building my startup Allocate, there was a time when I started feeling that I needed to better understand why my customers cared about some things and didn’t care about others. However, it wasn’t until I read Clayton Christensen’s Competing Against Luck that this idea hit home for me. After reading Competing Against Luck, I was surprised to find this concept in Don Norman’s The Design of Everyday Things, Simon Sinek’s Start With Why, and The Book of Why by Judea Pearl.
Christensen calls it the Theory of Jobs to Be Done; Norman talks about ‘root cause analysis’, the effectiveness of the 5 Whys by Sakichi Toyoda, and Design Thinking (which has a ‘why’ component to it); for Sinek, well, it’s the title and focus of his book; and Pearl describes the ‘new science of causal inference’. I don’t mean to say that all four authors are talking about the exact same ideas, because they’re not. There is, however, a central thread that connects them. That thread is, if you can understand why things happen, or the causal mechanism at work, then you are equipped to use this knowledge to do better work.
This idea likely seems obvious to you. It seemed obvious to me until I read the first example in Competing Against Luck, where the importance and nuance of the idea sunk in.
In Competing Against Luck, Christensen describes a fundamental problem with innovation processes at many companies, which is that the data used to make decisions takes the form of “this customer looks like that one” or “68 percent of customers say they prefer version A over version B”. His point here is that none of this type of data tells you why customers make their choices.
To figure out why, he proposes his Theory of Jobs to Be Done, whereby you should seek to understand the following: what job are customers ‘hiring’ a product to do for them? The insight / shift in thinking is that customers have progress they are trying to make in their lives, and to make progress, they will ‘hire’ or ‘fire’ products that facilitate or impede such progress. If you understand the progress they’re trying to make, you then uncover the why.
He kicks the book off with a fun but illustrative example. A fast-food chain had hired a friend of his to figure out how they could sell more milkshakes. The chain had initially asked customers how they could improve their milkshakes. Make them chewier? Chocolatier? After months of this type of investigation, the results were in: nothing changed.
So, Christensen says that they thought of approaching the question differently: “I wonder what job arises in people’s lives that causes them to come to this restaurant to ‘hire’ a milkshake?”, he said. After approaching the problem from this angle, they quickly learned that in the morning, customers were ‘hiring’ the milkshake to make their commutes to work more interesting and to keep them full until lunch. Thus, the competition wasn’t other chains’ milkshakes but doughnuts, bananas, bagels, and other breakfast options. In the afternoon, customers were often ‘hiring’ the milkshake for a different purpose - parents tended to buy milkshakes for their children after school to reward them and/or to bond with them.
As mentioned before, Customer Development can often produce unreliable insights as people are unreliable sources of information. Recall that humans don’t want to tell people that their ideas are bad and they will often ask for a faster horse rather than a car. In this fun example, we see one, potentially both, on display. Simply asking customers what they wanted didn’t produce meaningful data; instead, the researchers had to dig deeper to understand the core driver of their behaviors.
Here are a few more examples of how understanding why can lead to useful insights. I’ve intentionally chosen examples from different domains:
CRM Software: On the surface, most people would think a CRM’s value proposition is that it allows sales reps & managers to manage their sales processes - contact storage, contact logging, pipeline management, etc. While this is true, it’s not the whole story. For companies, the real value of a CRM is that they’re able to keep their salespeople’s most valuable asset — their rolodex. Now, when sales reps move on, their rolodex stays with the company that pays for the CRM software.
Dyson Vacuums: Sure, a Dyson vacuum cleaner is quite possibly the best suction device for cleaning your home - but is that really why people are willing to spend $500+ for one? I don’t think so. Like Apple and the iPhone, Dyson and their vacuum cleaners were founded by a visionary founder, viewed as the highest quality, easiest to use, most innovative product in the vacuum category - owning a Dyson isn’t just about how well it cleans your house, it’s about what it says about your taste for high quality and innovative products.
Wine clubs: What’s a wine club good for? Your quarterly shipment of wine, right? That’s only part of the story. After all, it’s easy to drive to Whole Foods or Trader Joe’s to pick up some wine if you want some — do you need it shipped to your house? Rather, one of the reasons that people belong to wine clubs is because that wine or winery tells them or reminds them of a story. Maybe it’s the beautiful grounds they were overwhelmed by during their trip to Napa Valley, the uncharacteristically unassuming tasting room that finally made them feel like they could get into wine, or perhaps they visited a small winery and the winemaker him/herself poured the tasting while describing everything about the winemaking process. They’re not just drinking wine, they’re remembering the story they were told about the harvest, the particularly hot year the region had, or the once-in-a-lifetime earthquake that made this year’s production one of the most limited (and unique) ever - and they’ll be able to tell their friends all about it.
A true understanding of why requires deeper analysis than usually expected, but yields powerful insights.
What People Don’t Do Matters (A Lot)
Before we discuss how understanding why can drive better outcomes, I’d like to first make a point about the importance of the absence of information. As NNTaleb says in his book Antifragile, “Don’t mistake the absence of evidence for the evidence of absence.”
I’ve expressed several times now the following idea, in more or less words:
“If one can, through Customer Development, understand why customers care or don’t care about a product’s value proposition, then one can build an actually robust business supported by an actually compelling value proposition.”
What I want to highlight is the bit “...or don’t care”. Just as the silver rule is known to be more robust than the golden rule, the absence of evidence [of information] does not mean there is evidence of absence [of information]. This idea, known as via negativa, can be a more powerful source of knowledge than confirmatory (via positiva) data. Thus, what people don’t do is just as important, and likely more important, than what they do.
In the Udacity example, what people didn’t do (graduate from online classes in large numbers) was the basis for Thrun’s pivot. Understanding why people didn’t do that action led Thrun to the insight that people ultimately care about education because it leads to employment.
Thus, you can’t only look at what people do or say. You also need to pay attention to what they don’t do and don’t say, to fully understand why they behave as they do.
How ‘Why’ Relates To My Journey
My startup Allocate was founded on the idea that time tracking for services-based professionals (ad agencies, lawyers, accounting, etc.) was broken and needed to be improved (if not completely gotten rid of) through machine learning / AI. We knew that if we solved this problem, it would be a wedge into the market, after which we could sell them other products like expenses and invoicing / payments, which would significantly expand the size of the opportunity we were pursuing. We entered into the Lean Launchpad at Stanford with this thesis, and by the end of the class, we had spoken to over 100 potential customers who largely validated our hypothesis that time tracking sucked and needed to be improved. If we could improve it, they would buy it, they said.
Amazing, right? Now we just needed to build this product and they would come.
As mentioned before, we were wrong.
What we failed to learn during the Lean Launchpad was why people cared about the vision we were selling to them. We thought that they wanted a suped-up, sexy, AI-powered tool to help them better understand their business. With these insights, they would run their businesses differently, they said.
What our professional service provider customers actually wanted, however, was to make more money, and to be able to justify it. That’s a big insight that quite drastically changed our value proposition, what we built, who our key partners were, how we marketed the product, and how we sold it.
Getting back to my main message, doing Customer Development through this lens of understanding why can meaningfully impact one’s journey and help them:
Build a good product
Orient functions of the business around the right activities.
Operate with more capital efficiency
Let’s walk through each one with some examples to make it less theoretical.
Build a Good Product
A good product helps users accomplish what they’re trying to accomplish in their lives.
At Allocate, once we finally understood why customers cared about our product’s value proposition - making more money and being able to justify it - we leaned into this and built features that solidified this value proposition.
Of course, we didn’t falsely create billable time for them, but we did make algorithmic and UI improvements to make sure that they captured all of their time spent working on client work. And because they cared about justifying the time spent, we gave them tools to easily export the data that underpinned the hours they spent doing client work.
All of these improvements led to us building a better product, which led to higher NPS and more referrals, which led to more customers and more data, which led to a better product, and the cycle repeated itself.
Orient Functions of the Business Around the Right Activities
At Allocate, the marketing website for our legal-focused product spoke directly to why lawyers cared about our solution — making sure they captured every billable hour, allowing them to make more money.
Additionally, when giving demos to potential clients, we communicated this value proposition throughout the demo so they were well aware that our product solved their pains. We understood why they were coming to speak to us well before we got on the phone with them, which increased our contact form sign-ups and demo-to-free-trial conversion rate massively (and subsequently, free-trial-to-conversion too).
One can imagine how this would impact any function within a business. Above I gave examples that impacted Sales & Marketing, but Customer Support, Partnerships, Biz Ops, Finance, and more are likely to be impacted as well.
Operate With More Capital Efficiency
I think this one will be quite obvious by now. If one can understand why their customers care or don’t care about their product, and do so quickly, then one can head in the right direction sooner, with fewer iterations, causing one to operate with far greater capital efficiency.
In the Udacity example, it took them nearly two years to figure out why customers cared about education. At my startup Allocate, it took us 18 months to figure out which customers to target, and how to sell to them. The startup ecosystem is littered with stories like this - of companies that took far longer than ideal to head in the right direction.
The graphic below by Lenny Rachitsky (from this post) shows this dynamic via the time it takes to go from company founding -> Product Market Fit (PMF). In his article, he says that the median time from idea to feeling PMF is 2 years. Notice that some exceptional companies (Airtable, Slack, Miro, Figma) took far longer to first feel PMF.
It’s not necessarily the case, but it is likely that during this period between inception and PMF, companies ‘pivot’ from one idea to another until they feel that things are working. The pivot is somewhat glorified in Silicon Valley, but in my opinion, it shouldn’t be. Yes, it’s important to pivot with new information, but the ideal state is that you never have to pivot.
If you can achieve this, by understanding why sooner, you’ll get to where you’re going faster, and spend less money along the way - the ideal outcome for entrepreneurs and their investors.
Wrapping Up
The idea that entrepreneurs should understand why someone cares about something is so seemingly obvious that it gets overlooked amongst the multitude of signals one is bombarded with while trying to build a company. If an entrepreneur can figure out the job or progress a customer is trying to accomplish, and keep asking why until they figure it out, then they will be moving in the right direction when it comes time to build their business.
Pair this mindset with the rigor and framework of Customer Development, and an entrepreneur will have a massive advantage in building an ultimately successful business.
As the title of Christensen’s book conveys, they’ll no longer be ‘competing against luck’ in their search for a repeatable, scalable business model.
An analogy that I like to use is comparing building a startup with gradient descent. An entrepreneur is constantly making small decisions until they find their optimal end-point. See the image below:
Doing Customer Development as traditionally described is a lot like stochastic gradient descent, the purple line. One has a hypothesis, tests it, and iterates, doing this repeatedly. As a result, their journey may be a bit noisy, like the purple line.
Meanwhile, doing Customer Development with the addition of understanding why will help to reduce the noise, making one’s startup journey look a lot more like the blue path - one with fewer wrong turns, less stress, and fewer bouts of wondering “what the *&!$ am I doing with my life!?”.
This is actually where the analogy seems to break down a bit. Batch gradient descent, while less noisy, is known to be slower to find the optimum end state than stochastic gradient descent because it looks at a larger dataset when computing each iteration. As the reader knows by now, I strongly believe that combining Customer Development with understanding why can not only lead to a smoother journey but a faster one too. While it may seem like the analogy breaks down here, the best companies often win by moving slower initially so they can go faster later. My classmate at Stanford, Ron Tidhar, found this dynamic at play amongst more performant companies during his PhD research.
In this post I’ve proposed that, as you evaluate different parts of your business, your learnings can be improved by understanding why. Couple this idea with Customer Development and you’ll have the right toolkit to build a good product, orient functions of the business around the right activities, and operate with more capital efficiency.
Thanks for reading. Please reach out with a comment! I’d love to hear from you.
A massive thank you to Ron Tidhar and Scott Sage for their help reviewing this post.