The Value of an Unpredictable Customer

A company in our portfolio spent nearly two years building an AI tool for college students.

It’s a difficult demographic to sell to. Practically, students are shitty customers. It’s a volatile demographic; they generally don’t have much disposable income, their needs and routines change constantly, and getting them to commit to another monthly subscription is a challenge. Even when the product is useful, that doesn’t necessarily translate into a great business.

The company had managed to attract around 30,000 users. That was meaningful progress, but it had taken considerable effort to get there. Customer acquisition was difficult, churn was a real concern, and the economics weren’t particularly attractive.

From an investor’s perspective, the problem looked fairly straightforward.

We encouraged the team to explore selling directly to universities. Instead of convincing thousands of students to pay individually, why not have an institution purchase the software and provide access to its students?

A university has a much larger budget, an interest in supporting its student population, and the potential to bring thousands of users onto the platform through one contract. The revenue could be more predictable, and the company wouldn’t have to fight for every individual subscription.

I still think that was reasonable advice.

But while the team was working through its institutional sales strategy, it also experimented with a different approach to consumer acquisition. It leaned into user-generated content (UGC), refined the audience it was targeting, and started reaching students through content that made sense for how they actually discovered products.

The results changed our understanding of the business.

The company went from roughly 30,000 users to more than 300,000 in 2 months. The original demographic hadn’t suddenly become wealthier, and the underlying product hadn’t been reinvented. What changed was the company’s ability to reach the right people.

We had been thinking about how to make the customer more economically attractive. The company found a much better way to acquire the customer it already wanted.

That experience made me reconsider what we really mean when we call a customer unpredictable.

The customer wasn’t necessarily the problem

Consumer businesses are difficult to underwrite because so much can change outside the company’s control.

Acquisition channels become more expensive. Consumer preferences shift. Customers lose interest. A product can be popular for a month and forgotten the next. Even strong early growth can disappear when the company stops spending on marketing.

Low-ticket software makes those problems especially visible. When customers only pay a small amount, there isn’t much room to spend heavily acquiring them. The company needs some combination of efficient distribution, repeat usage, strong retention, and enough lifetime value (LTV) to justify the economics.

When those numbers don’t work, it’s tempting to conclude that the customer is wrong for the business.

But there are several reasons a consumer company can struggle, and they require very different responses. It comes down to the basics:

Product → Do people actually want what we’re building?

Distribution → Can we consistently reach the people who want it?

Monetization → Are we charging the right person, in the right way?

Retention → Does the product remain valuable after the initial purchase?

These four questions seem basic. In practice, they’re surprisingly easy to confuse.

Poor acquisition can make a good product look unwanted. Weak monetization can make an engaged customer look worthless. Strong marketing can make a product with terrible retention look like it’s working.

And churn can be especially misleading if the company hasn’t separated customers who genuinely need the product from those who were never a good fit. Side note: That’s why testing your ICP time and time again is essential.

In our portfolio company’s case, the rapid growth following a distribution change suggested that the original acquisition approach had been a significant constraint.

That doesn’t eliminate the need to prove retention or healthy customer economics. Three hundred thousand users are only as valuable as the relationships and revenue the company can sustain, and there’s still work to be done.

But it changes the diagnosis.

We were trying to improve the business model before fully understanding whether the distribution model was working.

There’s value in reaching a difficult customer

The interesting part about college students is that almost everything making them difficult to monetize also makes them a valuable audience for other businesses.

They have limited budgets, but they’re forming preferences and adopting products that may remain relevant long after graduation. They’re entering financial relationships, preparing for careers, and making decisions across education, housing, technology, entertainment, and dozens of other categories.

And they’re notoriously difficult to reach through conventional channels.

The harder the climb, the better the view

A company that figures out how to consistently reach and engage that demographic may be creating value that doesn’t immediately appear in revenue per user.

The first layer is the transaction itself. How much does the customer pay, how often do they purchase, and how long do they stay?

But a customer relationship can extend well beyond that initial payment.

A company may develop the ability to introduce relevant adjacent products, distribute other services, or bring customers into relationships with businesses that have a much higher willingness to pay.

It may also learn something useful about the market. With the right permissions and privacy safeguards, sufficiently broad usage patterns can inform aggregate research or analytics for organizations trying to understand that audience. This isn’t a free pass to commercialize personal information, and data licensing isn’t appropriate for every business. But useful market insight can be another form of value.

There is also strategic value in market access.

An established company might be able to build comparable software without much difficulty. Recreating an engaged audience, efficient acquisition channel, and trusted relationship with a demographic it has struggled to reach can be considerably harder.

That can matter in an acquisition.

The buyer may be interested in the product and revenue, but it may also be purchasing a position in a market that would take years to build organically.

The value of a customer isn’t always limited to what that customer is willing to pay today.

The important caveat is that having an audience and having durable access to that audience are different things. A viral campaign can produce thousands of signups without creating any lasting advantage.

The value comes from being able to reach the right people repeatedly, serve them well, and maintain the relationship as their needs change.

B2B and B2C don’t have to compete

One thing our experience reinforced is that consumer and business-to-business models aren’t always opposing directions.

Sometimes both belong in the same company.

Healthcare is a familiar example (B2B2C). A patient uses a service, while an insurer, employer, or health plan may pay for some or all of it. The individual receiving the value isn’t necessarily the organization funding it.

The same arrangement can work in education.

A student uses the software. A university pays because better access, engagement, or student outcomes are valuable to the institution.

But the company doesn’t necessarily need to abandon direct consumer sales to pursue those contracts.

In fact, consumer adoption can make the institutional business stronger.

Consumer adoption → Evidence of demand → Institutional willingness to pay

An institution evaluating a product with demonstrated student usage has a different starting point than one being asked to purchase something students have never tried.

The relationship can work in reverse, too.

Institutional contracts → Broader access → More consumer adoption

A university contract can remove the payment barrier for students and put the product in front of people who might never have discovered it independently.

The two channels can reinforce one another.

There’s a related idea in enterprise software. Companies often run pilots before committing to larger deployments, partly because early usage can reveal value that a sales presentation can’t demonstrate. The initial contract may be small, but the relationship can establish whether the product deserves a much larger place inside the organization.

The pilot customer isn’t necessarily valuable because of the first invoice. They’re valuable because of what that initial relationship might prove or unlock.

Of course, consumer traction doesn’t guarantee an institutional sale. Administrators care about budgets, compliance, outcomes, and procurement requirements that individual students may never consider.

And institutional distribution doesn’t guarantee that students will actually use the product.

Both sides still have to work.

But a company that can acquire consumers directly while also monetizing through institutions has more options than one forced to choose a single model.

What are we actually underwriting?

I think this is where the original investment lesson becomes useful.

Investors tend to favor customers with clear budgets, predictable purchasing behavior, and measurable lifetime value. That preference is rational. It makes revenue easier to forecast and the business easier to evaluate.

I’ve made that argument myself, and I expect I’ll make it again.

But the danger is letting an understandable preference for predictable economics become an assumption that the underlying customer relationship isn’t valuable.

We had a company serving a difficult demographic with an unattractive acquisition history. The B2B opportunity looked like a sensible way to improve the business.

What we hadn’t fully accounted for was the possibility that the team could dramatically improve distribution without changing its primary customer.

Once that happened, the opportunity looked different.

The company wasn’t simply selling inexpensive software to students. It was developing a much stronger way to reach a large audience that plenty of other businesses would also like to serve.

That doesn’t mean every product should stay consumer-facing. It doesn’t mean growth matters more than retention, or that a large audience guarantees an attractive exit.

And it certainly doesn’t mean founders should ignore investor feedback.

Good investors can help identify structural problems that a great growth chart might temporarily hide.

But founders are closest to the experiments. They can see how customers respond, which messages resonate, where the product gets shared, and what changes when they adjust how it reaches the market.

Sometimes the investor’s best contribution is a useful challenge. Sometimes the founder’s best response is to prove that the challenge was aimed at the wrong problem.

What matters is being willing to update the view when the evidence changes.

The relationship is the asset

I still like B2B models for many of the same reasons I did before. Larger contracts, concentrated purchasing power, and more predictable revenue can create excellent businesses.

But I think about difficult consumer segments differently now.

A customer with limited willingness to pay may still be worth serving. An audience with unpredictable behavior may still be valuable to understand. And a company whose initial subscription economics look unattractive may be building a distribution advantage that eventually supports several different revenue models.

The question is whether that advantage can endure.

Can the company continue reaching customers efficiently? Do those customers receive enough value to stay? Can the business convert the relationship into sustainable revenue, whether directly, through institutions, or through another model that genuinely serves the audience?

Those questions matter more than whether the customer fits an investor’s preferred profile.

In our case, the most useful lesson wasn’t that B2C beat B2B. The institutional opportunity remains worth pursuing, and the two could ultimately make each other stronger.

It was that we had underestimated what the existing customer relationship could become once the company figured out how to distribute the product.

Sometimes a customer looks unpredictable because the market is genuinely difficult.

Other times, the business just hasn’t learned how to reach them yet.

The challenge is knowing which problem you’re actually trying to solve.