Skip to content
Aplora
EdTech
Aplora Sales

We validated the system inside our own operations first

The funnel from lead to purchase in one picture: where the money was actually leaking

An online IT school in Poland with a multi-step funnel: ads → lead → webinar signup → attendance → call → second contact → purchase. Marketing and sales each analysed their own stretch of it.

Published: 2026-08-24 · updated: 2026-08-24

A tablet showing an analytics dashboard: a pie chart and a trend line
Illustrative photo of the working context, not a screenshot of the client's system

  • Google versus Facebook lead quality by conversion to sale

    ×3 in Google's favour

  • Conversion at four or more touches versus one

    ×4–5

  • Potential estimate: sales from the hot-lead pool identified

    +14–22 sales

In short

  • Problem: marketing and sales looked at different halves of one funnel, and the losses between them were invisible to everyone.
  • Solution: joining the funnel into a single chain and analysing it by channel, segment and touch count, with the eventual sale as the metric.
  • Result: the quality gap between channels, the effect of touch count on conversion, and a specific pool of leads worth calling first.
  • Systems: ad platforms, webinar platform, quizzes, CRM.

Context

  • An online IT school selling B2C through a webinar and a consultation call.
  • Team: marketing and sales, working in different systems and off different reports.
  • Process volume: the full chain lead → webinar → call → second contact → purchase.
  • Systems: ad platforms, webinar platform, quizzes, CRM; part of the data under NDA.
  • The constraint became visible when raising the ad budget stopped producing a proportional rise in sales.

Baseline

Baseline metrics with their sources. Without them, any later result has nothing to be compared against.
  • Channels are compared on cost per lead, not on the sale
  • A cheap channel wins by bringing people who don't buy
  • Nobody counts the losses between stages
  • Marketing and sales draw the funnel differently
  • Before the review we fixed the actual conversion at each funnel stage, per channel and per number of touches before purchase.
  • Data source: exports from the ad platforms, the webinar platform and the CRM covering the same period.
  • The comparison ran on the eventual sale, not on clicks or leads: substituting that metric was exactly what made the cheap channel look attractive.
  • There are no measurement dates or calculation formula beside the numbers, so they are published as reported results.

Diagnosis

Which hypotheses were considered, why this one was chosen, what was assumed, and the condition under which we would have stopped.
  • Three hypotheses were on the table: too little traffic, a weak offer, or losses inside the funnel.
  • The third was chosen: lead volume was growing and sales were not, and the gap sat in the stretches between systems that nobody measured end to end.
  • The assumption: last-touch attribution distorts the picture in a multi-step funnel. It was tested by recalculating across the whole touch chain.
  • Stop criterion: if, once the chain is assembled, the difference between channels falls within statistical noise, no budget conclusion is drawn.

What we implemented

  • This is a diagnosis, not an implementation: the output is a picture of the funnel and a plan, not a running system.
  • Data sources: ad platforms, webinar platform, quizzes, CRM — joined on a single lead identifier.
  • What was analysed: stage-by-stage conversion, lead quality per channel measured against the eventual sale, the effect of touch count, and segments from quiz answers.
  • What the client received: a list of decisions with an expected effect for each — move budget, extend sequences to four touches, call the identified pool.
  • Human checkpoints: the head of the funnel sets the priority, not the calculation: some conclusions run into constraints the data cannot see.

How the process changed

Before

5 steps
  1. Marketing measures cost per lead by channel
  2. Sales measure conversion from call to deal
  3. Reports are merged by hand and over different periods
  4. The budget decision is made on cost per lead
  5. Losses between the webinar and the call go unmeasured

After

5 steps
  1. Data is joined on a single lead identifier
  2. The funnel is measured end to end: lead → webinar → call → second contact → purchase
  3. Channels are compared on the eventual sale
  4. Segments and touch counts are analysed separately
  5. The output is decisions with an expected effect

What was stuck

Channels were compared on cost per lead rather than on the eventual sale, so a cheap channel looked better than an expensive one even when it brought people who don't buy. Nobody counted the losses between stages: for marketing the funnel ended at the webinar, for sales it began at the call.

  1. 1Data is joined on a single lead identifier
  2. 2The funnel is measured end to end: lead → webinar → call → second contact → purchase
  3. 3Channels are compared on the eventual sale
  4. 4Segments and touch counts are analysed separately
  5. 5The output is decisions with an expected effect
  6. Measured result

Results

Google versus Facebook lead quality by conversion to sale

×3 in Google's favour

Reported by: An observation from LearnIT's own funnel data over the period reviewed

Conversion at four or more touches versus one

×4–5

Reported by: An observation from LearnIT's own funnel data over the period reviewed

Potential estimate: sales from the hot-lead pool identified

+14–22 sales

Reported by: Aplora's calculation from a comparable segment's conversion: an estimate of the potential from calling 57 hot leads, not a measured result

Economic impact

  • The effect comes from redistributing budget already spent rather than adding to it: the same outlay, a different split across channels.
  • The potential estimate from calling the hot pool is a projection based on a comparable segment's conversion, not an outcome achieved. The difference matters, and the page says so.
  • Cost of implementation is the analysis work; a diagnosis carries no ongoing TCO, but it holds no effect either: without process change the picture goes stale within one intake.

Adoption

  • Marketing and sales started looking at the same funnel — until then each team had its own.
  • Budget decisions moved onto the eventual sale rather than cost per lead.
  • The solution owner is the head of the function: a review with no owner becomes a deck nobody opens twice.
«For the first time we saw the whole funnel: it became clear where we lose money and what to actually do — where to move budget, whom to follow up and which touches genuinely produce sales.»

Client under NDA — Head of sales and growth, LearnIT

What's next

  • Scaling: the same review on the next intake, to tell a durable channel difference from a one-off.
  • Next initiative: lead scoring, so the pool surfaces automatically rather than through a periodic review.
  • What we decided against: dropping the weaker channel outright. A quality gap does not mean the channel is useless: it brings a different segment, and that was worth testing separately on the next pass.

Have a similar workflow? Let's check whether the hypothesis transfers

Another company's result is not a promise. It does show where to look.

All cases