“Data driven” decisions aren’t innovative decisions
Focus Groups. Surveys. Segmenting our customer data. Segmenting competitors’ customer data. Digging through people’s smartphone data.
The modern tech product company wants to be data driven—and they’ll do whatever they can to get their hands on that data.
Compare that to the approach that produced the unicorn companies of the late nineties: your eBays and your Amazons, for example. The archetypical founding story? Some guy—admittedly it’s usually some wealthy, well-connected guy, but some guy nonetheless—thought up a thing that he wanted for himself. It turned out that the world wanted it, too, to the tune of billions of dollars.
Why aren’t the unicorn companies of 2023 talking about how some dude in his garage wanted a thing for himself in 2013?
The answer derives from who built the internet of today, and who they built it for.
Innovation serves unmet needs
An innovative product breaks the market for existing solutions. And this is important: it usually doesn’t just serve a need better than other solutions serve it. Instead, it serves a need that other solutions don’t serve at all.
Let’s look at some examples.
eBay used reputational scores to connect buyers and sellers who did not know each other and have stuff actually get paid for and delivered—across time and space! Until the early internet, sales had to be colocated, and even after the internet they carried the risk of some faceless website stealing your money and never sending the thing you ordered. At the time, no mainstream solution did this.
Amazon added a logistical system that made ordering things over the internet not only reliable, but also relatively fast. No more waiting three weeks for John in Ohio to ship your CrazyTown album sticker via snail mail. To this day, for better or worse, Amazon offers a more streamlined, reliable online purchasing infrastructure for many products than even the sites of the shops that create the things they’re selling.
When Apple released the iPhone in 2007, no other phone offered the opportunity for different apps to have different input interfaces. Its introduction more or less broke the market for existing phones.
In the days of the early internet, unserved needs were easy to find: until then, commerce had to happen either in person or via snail mail and phonecalls. Quick, self-service, asynchronous options simply did not exist. The web made it possible for them to exist. It more or less dumped a whole vat of krill into the waters of commercial demand for companies to snap up without having to be that creative.
Now, those krill have largely been eaten. The easy wins have been won. Online commerce is a saturated business. The commercial web’s early garage founders built the things they wanted, and now the things that that demographic wants have been built.
Focus groups, surveys, segmented customer data, segmented competitor customer data, smartphone data—today’s product and marketing teams lean heavily on these resources to find the tiniest capillary opportunities for monetization within that demographic.
Innovation means, pragmatically, serving a heretofore unserved need. Those are hard to find among the people that the internet was built for. The problem lies in the fact that the product and marketing teams’ sources of data—people who do focus groups and fill out surveys from their smartphones, people who already use the product or a competitor’s product or a variety of smartphone apps—come from the people that the internet was built for. These people have few unserved needs online.
In search of the unserved need
So where can we find the unserved needs—the opportunities to innovate? It will sound too basic to be true when I say it: we have to talk to folks that don’t enjoy an internet built for them. We can’t find them among folks already well-represented in boardrooms, product teams, and pools of active customers. Instead, visionary products derive directly from centering people at the margins of modern technology.
That doesn’t sound like it would be a profitable strategy, does it? Our instinct on this tends to get hindered by two flawed assumptions.
Flawed Assumption #1: “Marginalized folks constitute a minority of our potential customers.”
This is just flatly incorrect in a lot of cases. Take a look at the difference in uptake by race for online money transfer companies CashApp and Venmo. These two companies do not differ in size by orders of magnitude. A Venmo employee might say of the company’s numbers “That’s just who wants/needs a money transfer app.” And they’d be, candidly, wrong.
Flawed Assumption #2: “Pursuing solutions in underserved populations will cost us demand from existing customers or larger demographics.”
This would be true if the needs of the two groups were mutually exclusive, but they’re often complementary. My favorite example of this is the same 2007 iPhone introduction I mentioned above. The star feature of that phone—the multi-touch capacitative screen that facilitates different interfaces for different apps—was not an Apple invention. Apple got that screen by acquiring FingerWorks, a small company founded by John Elias and Wayne Westerman to distribute a computer input device Westerman had designed to help his mother, who lost much of her fine motor function, to continue to do the activities she wanted to do on her computer. Indeed, you’d be floored by the number of your favorite technological features that started as accessibility features for people with disabilities.
Now, reams have been written about how to get marginalized people into the room on product decisions. Tech companies individually tend to experience intermittent hiccups of trying really hard on that, but as an industry as a whole, the situation ain’t great. No number of DEI trainings seems to manage to keep a diverse team. And I maintain that the reason for this is that, as not-racist, not-sexist, or not-ableist as individuals want to be, we run into two things:
- There is so much to learn about so many axes of marginalization, that a single corporate training could not possibly completely inform and educate about all of them.
- An equitable environment that welcomes diversity requires its members to have a complement of skills, but most training on the matter focuses on values or demographic groups rather than skills, leaving folks unequipped to uncover and examine unserved and underserved needs that technology could address.
Teams can facilitate an inclusive environment and increase their capacity to sniff out unserved needs by explicitly practicing that complement of skills in their work. Some of these get touched on in DEI trainings, but they form the foundation of creating a team where everyone experiences positive changes in their opportunity to contribute. I have written before about building an employee evaluation rubric for what I see as the critical five:
This is the skill of giving folks in a meeting the opportunity to listen by protecting their opportunities to speak. Without this, teams run into all kinds of icky demographic patterns governing who gets to talk in meetings, and it’s often not the people with the most thought-out ideas. That’s why moderation is such a critical skill—especially for leadership of an innovative team. You need people’s ideas, and that requires the skill of uncovering them in groups.
2. Soliciting opinions
There’s also skill associated with uncovering ideas individually, and identifying who to ask for those ideas. It includes thinking specifically about who will be affected by a decision, and seeking them out for input. It also includes identifying which perspectives would disagree with those making the decision, and explicitly soliciting those objections to avoid running into obstacles later on that could have been predicted.
This often feels like a small skill, but it has a huge impact. It’s common in tech for folks to misattribute ideas, forget where they heard them, or fail to recall who taught them something they know. In the aggregate, this makes folks less likely to share their ideas out of concern that they’ll lose credit for them. It’s especially true among underrepresented groups, but it can happen to anyone. A team that cares about attribution, and practices making it regularly, facilitates earlier and more open idea sharing.
4. Most advanced assumption
It’s easy to accidentally condescend to others in the workplace and give the impression that we think other people are less capable than we are. We want to share what we know, and sometimes the unintentional result of that is that we lecture someone who knows more than we do. This makes folks feel like we don’t respect their valuable expertise. And if we don’t respect it, why should they share it?
This is another one that feels small but makes an outsize impact. It’s also tricky to broach a topic when we don’t know the level of expertise of others in the conversation. But we can learn to ask questions about that. We can also learn to share knowledge with explicit consent to be stopped if an interlocutor already knows, or if we assume too much knowledge and they have questions.
5. Capitalizing on alternative perspectives
We’re a conflict-avoidant culture: we’ll go to great lengths to skirt disagreement. But disagreement often signals and precedes conversations that result in more complete solutions that better serve unserved needs. Developing a more positive attitude about disagreement requires tools for navigating it more comfortably and with more grace for each other. And that, like the other four items discussed above, is a skill that can be taught, learned, and practiced.
Those skills all have a lot of nuance in them, and I have not explained exactly where to get it. I’m working on an online, self-paced workshop with exercises in it to help technologists like you create and lead teams that can, in fact, innovate. You can pre-order that course right here (and, if you want proof of my teaching chops, this course on tech debt is already up and available).
But the takeaway here revolves around where innovation comes from. We can employ skills—individual, interpersonal skills—to create the kind of environment that makes innovation possible, even in a tech sector that often looks bleakly saturated. There’s hope for us, hope for our teams, and hope for our industry if we can learn to execute on our curiosity about the unmet needs we can build for.Tags: data science, software development
I don’t completely disagree, but I do feel you’re throwing out the baby, bassinet, diaper pail etc etc with the bathwater.
Data DOES matter. Sure the next great thing might not come from looking at all the current things, but shitty customer service, yeah, that comes from all the current things.
So… tone it down, because you’re half right.
I feel like you’re mixing things up here. I don’t think innovation/disruption now happens significantly differently or less than it did 10, 20 or more years ago. Sure, there is a big emphasis on data now, and, sure, it’s over-hyped, but that is not stopping innovation. It might just mean that some people, that aren’t otherwise innovative, are trying to find something using a process that they think might work for them. In the end, even with data, innovation still happens the same – someone takes what they know and develops a unique idea and moves forward with that. Data always played into that process, but so does the ability to envision new things. If it feels like that is not happening as much as it used to, I suspect that is due primarily to the fact that real innovation like what you’re using as examples is relatively rare and there is a difference in perspective between looking at current events vs looking at history. Over broader timelines, if feels like a lot of innovation has happened. But looking at the near term, we don’t see it, because we’re overwhelmed with everything we see now, vs everything we’ve filtered out of our historical perspective, which leaves just the highlights for us to consider for the past. I suspect that 10 years from now, we’ll be looking back that this period like we look at 10 years ago, and it won’t look like it does to us now.