Ilya Rudakov
← All cases

Monetizing the user base

How we shortened the repeat deal cycle and grew repeat purchases from 7% to 33% of revenue: a short format for alumni and add-on products

The mini app in the Telegram bot: home screen, buying art-director reviews and extending access

7% → 33%

repeat purchases as a share of revenue: April 2025 → October 2026

3–⁠4 wks

repeat deal cycle instead of ~12 months: from graduation to the next purchase

+26%

revenue from add-ons year over year

Context and goal

Product and metrics

A base of 85K users; the core product is 3–⁠4-month intensives with a heavy workload. In April 2025 repeat purchases made up 7% of revenue, and the repeat deal cycle — from graduation to the next purchase — was about 12 months.

Causes and goal

After a demanding intensive, demand for the next big product dropped: alumni needed a rest. Add-ons — access extensions and work reviews by an art director — barely sold, even though satisfaction with the courses was high. The goal: shorten the repeat deal cycle and grow the share of repeat revenue.

My role

I led the repeat-purchase area: the shorts format, their pages and the path to payment, the add-on research and the decisions that followed. Cohort communications were a whole marketing-team effort: together we segmented alumni and decided what to offer whom and when, and an editor wrote the copy.

Approach: a next step, at the right time, with no extra effort

A repeat purchase happens when the customer has a manageable next step, the offer arrives at the right moment, and buying takes no messaging or searching. Each decision covered one of these three conditions.

  1. 01

    Next step

    A product alumni are ready for after a demanding intensive.

    Shorts: a short format at a comfortable pace

  2. 02

    At the right time

    The offer arrives when the customer needs it.

    Bot reminders before access ends, cohort communications

  3. 03

    Easy to buy

    Buying happens where the customer already studies.

    A mini app in the Telegram bot instead of messaging support

The approach works in any product where repeat purchases matter: subscriptions, upsells, services with a long customer lifecycle.

01

Shorts: the next purchase in weeks rather than a year

The main lever of repeat revenue is the time from graduation to the next purchase. For us it was about 12 months.

Why alumni didn’t come back. After 3–⁠4 months of heavy workload they needed a rest, not another big course. The demand was there, but it was put off for almost a year.

A bet on a short format. In May 2025 we launched shorts: 2–⁠4-week intensives on one skill — video lessons, expert feedback several times a week and a portfolio piece every week. Each topic ran once, so the purchase had a real deadline with no fake discounts.

What came of it. Shorts were bought mostly by alumni: repeat sales went up, and the repeat deal cycle shrank from about 12 months to 3–⁠4 weeks. Together with add-ons, repeat purchases grew from 7% to 33% of revenue.

The Design System in Claude Code shorts page on desktop and mobile
Design System in Claude Code: “we run it only once, no repeats”
The AI Brand Identity Concept shorts page
AI Brand Identity Concept: the last run on this topic
› See the whole Post AI Creative & Motion page scroll
› See the whole Design System in Claude Code page scroll
› The full AI Brand Identity Concept page
The full AI Brand Identity Concept shorts page

02

Add-ons: demand without purchases

The other part of repeat revenue is add-ons to what customers already bought: access extensions and work reviews by an art director.

What was unclear. Add-ons barely sold, and they had never been tested with users. It wasn’t clear whether there was no demand or the offer simply didn’t reach the customer at the right moment.

How we checked. I ran 20 in-depth interviews with alumni across four buyer segments — people who had bought an extension, a work review or both — and asked separately what price they consider fair.

What we found. The demand is there: half extended access to finish a portfolio project. And about half of the buyers didn’t know how to buy and messaged a manager directly.

50%

Hadn’t finished the course

The main reason to extend: get to the end and leave with a finished portfolio project.

≈50%

Didn’t know how to buy

People learned about the option from chats. After the course they wrote to support or to a manager directly.

2

Review levels

Some need a quick direction in 15 minutes, others a detailed review with questions.

«When I extend, I expect the whole package to be extended. But here everything is separate.»
«It was important for me to do a small recap so I could feel calm at interviews.»
«My feedback sessions expired. I paid for a service and didn’t get it.»

03

The offer at the right moment and buying in the bot

If demand exists but purchases don’t happen, the fix is the timing of the offer and the path to payment.

A reminder before access ends. At first we reminded people by hand, then moved the reminders into the bot: before access ends, a message explains why it’s worth extending and has a buy button.

Buying where people study. In the bot’s mini app, lessons, feedback booking and the “extended study” add-ons section sit on one screen: a 3, 6 or 12-month extension and a pack of 3, 5 or 7 reviews are paid for without messaging support.

What came of it. Add-on revenue grew by 26% year over year.

The result. Repeat purchases grew from 7% of revenue in April 2025 to 33% by October 2026: the repeat deal cycle shrank from about 12 months to 3–⁠4 weeks, and add-ons started to sell.

Which hypotheses worked

  1. 01

    If you offer a short course after a big one, alumni come back sooner. The repeat deal cycle shrank from about 12 months to 3–⁠4 weeks, and together with add-ons repeat purchases grew from 7% to 33% of revenue.

  2. 02

    If you remind people about extending before access ends, they buy it. Add-on revenue grew by 26% year over year.

  3. 03

    If you simplify buying — remove extra steps and messaging support — people buy more. Extensions and reviews are paid for in the bot’s mini app, right where people study.

  4. 04

    If it’s unclear why add-ons don’t sell, talk to buyers from different segments. 20 interviews across four segments showed that demand exists, and half of the buyers didn’t know how to buy.

«I recommend Ilya without reservation as someone who will get to the bottom of anything and find the shortest and most effective path to solving a problem.»

Maria Teplenko · CMO of wannabe & Roots