When the user wanted the product but could not be the customer

The startup had a product people genuinely needed.

Across an illustrative six-month operating period, roughly 1,400 learners registered for the platform, around 520 became meaningfully active, and about 160 households started a paid subscription.

At approximately $39 CAD per month, that should have been encouraging.

It was not.

Paid conversion was only around 11% of registered users, and many families who engaged with the product did not remain paying customers long enough to create sustainable recurring revenue. The uncomfortable part was that usage data did not suggest the product was simply unwanted.

Some of the people who appeared to need it most were among the least able to pay for it.

That changed the question we were trying to answer.

The issue was no longer, "How do we persuade more users to subscribe?"

It was, "Are we asking the right person to pay?"

The product was solving a real problem

The company had developed educational software to help people with disabilities improve their English reading and writing skills.

The need was easy to understand. Better literacy could support school participation, communication, employment readiness and greater independence in everyday activities.

The original business model was also straightforward.

A learner or family discovered the product, registered, tried it and, if they found it useful, paid for continued access.

In a conventional consumer education business, that funnel could make sense.

But this was not a conventional consumer education market.

Using a reconstructed operating model, the company might generate roughly 230 new registrations per month. About 85 users would become meaningfully active, perhaps completing three or more learning sessions. Yet only around 25 to 30 households would become paying subscribers.

That meant engagement among interested users could be several times higher than paid conversion.

The first assumption was that something in the funnel was failing.

Perhaps the price was wrong. Perhaps onboarding was too difficult. Perhaps marketing was attracting the wrong people.

Those were reasonable possibilities.

But they did not explain the entire pattern.

A $39 subscription was not really a $39 decision

For many families, the subscription existed alongside other expenses.

Therapy, specialized tutoring, transportation, assistive technology, caregiving and other forms of support could all compete for the same household budget.

In our working affordability model, we considered a family already allocating roughly $400 to $900 CAD per month to a combination of disability-related services and support.

Against that background, another $39 CAD monthly subscription, or nearly $470 CAD annually, was not necessarily a trivial purchase.

Reducing the price might improve conversion slightly.

But the economics quickly became uncomfortable.

Suppose the company reduced the subscription from $39 to $25 CAD per month and improved paid conversion from roughly 11% to 15%.

With 230 new registrations per month, that would mean approximately 35 new paying households rather than 25.

The company would gain around 10 additional customers, but each customer would contribute $14 less revenue per month.

The pricing change could make access easier without fundamentally solving the business problem.

More importantly, it would still leave the financial burden with the family.

That was the first important shift in our diagnosis.

The person receiving the value did not necessarily have the financial capacity to become the customer.

Affordability was only one part of the problem

As we examined the experience more closely, another assumption began to break down.

"People with disabilities" could not be treated as one homogeneous user segment.

A learner with dyslexia, a learner with low vision and a learner with limited motor control might all be working toward better literacy, but the interface and learning experience required by each could be substantially different.

The same problem appeared with age.

A 16-year-old learner might have a reading level closer to that typically associated with an eight-year-old. Giving that learner content designed visually and culturally for an eight-year-old might technically match the reading difficulty while completely missing the emotional and motivational requirements.

Session length created another constraint.

A conventional educational product might optimize for a 25 to 30 minute learning session. For some users in this population, a more realistic productive session might be 10 to 15 minutes, followed by a break.

That changes product design.

Instead of measuring success primarily through lessons completed or time spent in the application, the system needed to accommodate different combinations of cognitive load, reading ability, motor access, vision, communication method and energy.

For some learners, text-to-speech might be central.

For others, speech-to-text could matter more.

Some might need slower progression, shorter exercises, larger interface elements or a different page layout.

The challenge had moved beyond teaching English.

We were designing around the question:

How can a learning path adapt to a person's ability, limitations, age, energy and real-life environment?

The learner was not using the product alone

There was another stakeholder in the system.

In many cases, the learner's experience depended partly on a parent, caregiver, teacher or therapist.

That created an important operational problem.

Even if a learner completed three supervised sessions during the week, the difference between continuing practice at home and abandoning the product could depend on whether the person supporting them understood what to do next.

Traditional onboarding focused almost entirely on the learner.

We concluded that the support system around the learner required its own experience.

Instead of expecting a parent to interpret detailed educational data, the product could translate progress into simple actions.

For example, after a week of three or four short sessions, a caregiver might receive one practical recommendation for the following week, together with a simple explanation of what the learner was working on.

The objective was not to turn the caregiver into a teacher or therapist.

It was to make support easier to provide.

That changed the product architecture again.

There was no longer one user experience.

There were several connected experiences around the same learning outcome.

Then the business model problem became visible

By this stage, three different roles had emerged.

The user was the person interacting with the learning product.

The beneficiary included the learner and family, but potentially also health, education and social-support systems that benefit when a person becomes more capable and independent.

The payer, however, was still assumed to be the family.

Those roles did not align.

A simple numerical example showed why that mattered.

Imagine 100 families with a genuine need for the product.

Perhaps 60 would try it.

Around 35 might use it enough to demonstrate meaningful interest.

But if only 12 could justify another $39 CAD monthly expense, the business would monetize just a small portion of the population receiving potential value.

The company could interpret that as a conversion problem.

We saw something different.

It was a payer-market-fit problem.

The next question therefore became more useful than another pricing experiment:

Which organization benefits economically when these individuals achieve better functional outcomes, and could that organization fund access instead?

From EdTech subscription to fundable intervention

One potential answer was an insurer or another institutional payer.

But replacing a credit-card form with an insurance invoice would not create a viable B2B2C model.

The product itself had to become easier for a payer to understand and evaluate.

A family could reasonably describe the platform as a learning tool.

A payer would need different answers.

Who is eligible?

What specific need is being addressed?

What intervention is being delivered?

How long does it last?

How is improvement measured?

What evidence shows that the member is progressing?

And eventually, can improved capability plausibly contribute to better functional outcomes or reduced dependency?

We therefore reframed the product from a general educational subscription into a measurable accessibility and learning intervention.

A possible program structure was a 12-week initial intervention, rather than an indefinite consumer subscription.

A learner could complete a baseline assessment, receive an individualized learning path and work through approximately three short sessions per week.

At an average of 15 to 20 minutes per session, that represents roughly 12 hours of structured learning activity over 12 weeks.

Progress could then be reviewed at baseline, approximately week six and at the end of the initial program.

The product had become much easier to describe operationally.

More importantly, it had become easier to fund.

Measuring something more useful than lessons completed

Consumer software commonly measures engagement.

Lessons completed.

Minutes used.

Weekly active users.

Those numbers still mattered operationally, but they were not enough for the redesigned model.

A payer needed to understand whether access was contributing to meaningful improvement.

We therefore shifted the outcome model toward measures such as reading comprehension, practical writing ability and the learner's capacity to complete relevant activities with less assistance.

For example, a learner might begin the program requiring support on 8 out of 10 reading-related daily tasks identified in the assessment.

After a structured intervention, the relevant question would not simply be whether they had completed 40 lessons.

The more meaningful question might be whether the learner now required support on five tasks instead of eight.

That is a very different definition of product value.

It connects product activity to functional improvement.

The measurement model would still need clinical and payer validation before reimbursement claims could be made. But strategically, the direction was clear.

The company needed to measure the outcome the funding organization cared about, not merely the activity the software could most easily track.

The administrative layer became part of the product

A consumer subscription can begin with an email address and a credit card.

A funded intervention cannot.

Eligibility may need to be established.

A referral may be required.

Authorization may need to be recorded.

Assessment documentation has to be stored.

Progress has to be reported.

A renewal decision may depend on evidence.

This meant the company was no longer designing only a learning application.

It also needed an administrative layer capable of supporting the funding workflow.

The business architecture was becoming:

Payer → Platform → Member

with caregivers and relevant professionals participating around the member where appropriate.

The complexity increased, but something else improved dramatically.

The identity of the customer finally began to match the economics of the problem.

What the new economics could look like

The difference becomes clearer in a simple commercial model.

Under the original consumer model, 160 paying households at $39 CAD per month would produce approximately $6,200 CAD in monthly recurring revenue.

To reach $25,000 CAD in monthly recurring revenue at the same price, the company would need roughly 640 paying households.

If only around 11% of registered users converted, generating that customer base could require thousands of qualified consumer registrations, followed by continuous effort to replace cancellations.

The institutional model changes the unit of sale.

Suppose a payer made the intervention available to 500 eligible members.

If approximately 35% enrolled, the program would serve around 175 active members.

At a modeled reimbursement of approximately $65 CAD per active member per month, that population would represent roughly $11,000 CAD in monthly program revenue, or approximately $135,000 CAD annually if participation remained broadly stable.

Two payer relationships of similar size could therefore generate revenue comparable to several hundred individual consumer subscriptions.

The point was not that institutional funding automatically produced better margins.

It came with new costs.

Assessment, member support, reporting, documentation and payer administration might raise direct service cost to approximately $25 to $30 CAD per active member per month, compared with the much lower marginal cost of distributing conventional software.

At $65 CAD of monthly reimbursement and roughly $28 CAD of direct program cost, the contribution before company overhead would be about $37 CAD per active member.

For 175 active members, that would be roughly $6,500 CAD per month of contribution before fixed operating costs.

The model still had to prove retention, outcomes and acquisition economics.

But it created something the original model struggled to provide: a plausible mechanism for separating access from household purchasing power.

The most important change was the definition of value

The original value proposition could be summarized as:

"We help people with disabilities learn English."

That described the product.

It did not fully describe the business.

The redesigned proposition moved closer to:

"We help people with disabilities build functional literacy and greater independence, while giving payers a measurable way to fund and track that improvement."

This was not simply new marketing language.

It changed what had to be built, measured and sold.

The learner needed an accessible, adaptive learning experience.

The caregiver needed understandable guidance.

A professional might need visibility into progress.

The payer needed eligibility logic, documentation and outcomes reporting.

The company was no longer designing one software interface for one customer.

It was designing a system connecting several stakeholders around one measurable outcome.

Before and after the strategic reframing

Original model	Redesigned model
—	—	—
Primary buyer	Individual household	Institutional payer
Illustrative consumer price	$39 CAD per month	About $65 CAD per active member per month
Unit of sale	One subscription	Covered member population
Illustrative paid users	About 160 households	About 175 members from one 500-member eligible population
Monthly revenue example	About $6,200 CAD	About $11,000 CAD
Core product metric	Engagement and lesson completion	Engagement plus functional outcomes
Onboarding	Learner-focused	Learner, caregiver and administrative workflow
Primary value	Learning English	Functional literacy, accessibility and measurable progress
Commercial constraint	Household affordability	Ability to demonstrate fundable outcomes

These figures were useful not because they predicted the company's future precisely, but because they showed how changing the payer could change the mechanics of the business.

Product-market fit was not enough

The central lesson from the project was that Product-Market Fit without Payer-Market Fit can still produce a fragile business.

A company can solve a genuine problem.

Users can want the product.

Engagement can be real.

The product can even produce meaningful value.

None of those conditions guarantee that the person receiving the value should also be the person expected to pay for it.

That distinction becomes particularly important in markets where the end user has significant needs but limited discretionary purchasing power.

In those situations, weak conversion is not automatically evidence of weak demand.

Sometimes it is evidence that the business has assigned the role of "customer" to the wrong participant in the system.

The strategic question that changed this case was therefore not:

How do we extract more revenue from the people who need the product?

It was:

How might we make the people who benefit from the product different from the people who have to pay for it?

Once we asked that question, the company was no longer dealing with an ordinary EdTech growth problem.

It was dealing with accessibility, outcome measurement, healthcare financing and system design.

And that was a much more useful problem to solve.

Questions for other businesses

• Is the person using your product actually the stakeholder with the strongest ability and incentive to pay for it?
• If customers repeatedly engage but fail to convert, could affordability or payer structure explain the gap better than marketing?
• Which organization captures economic value when your customer achieves a better outcome?
• What would you need to measure for that organization to justify funding your product?
• If you moved from a direct-to-consumer model to B2B2C, what new administrative capabilities would become part of the product?

For businesses serving populations where need and purchasing power do not naturally align, improving the conversion funnel may be the wrong starting point. Sometimes the more important task is redesigning who pays, what they are paying for, and how the value created for them can be measured.