USD 1.23M and 85,275 applications from a systematic acquisition department

In May 2022, an online school for fitness and nutrition professionals had lost stability after depending too heavily on one social platform. Specialists were tested reactively, end-to-end reporting was unreliable, and the founders were coordinating marketing while running the rest of the company.

A website application cost more than RUB 900, about USD 10.68 at the reference rate used on this page. May closed at roughly 150% ROMI, below what the business model required. Some vendors produced applications at RUB 8,000, approximately USD 94.92. The larger problem was diagnostic: the company could not quickly separate sources that produced sales from sources that only filled a dashboard.

• Dependence on one demand source
• Vendor evaluation without consistent revenue attribution
• No owner of the complete acquisition system
• Volume targets disconnected from profitability

When everyone buys traffic but nobody owns the economics, doubling the budget only accelerates the errors.

The first move was an audit of every active specialist and campaign. Combinations with major setup errors or no commercial evidence were removed. Only sources already connected to sales, or showing a verifiably sound setup, remained.

Spend, source, application, payment and revenue were then connected in one end-to-end view. The integration kept evolving as new decisions required deeper fields. The team moved to a weekly cadence: review data, agree changes, record hypotheses and give campaigns enough time to learn before touching them again.

Between 00:00 and 00:59, RUB 539,532 in spend had accumulated — about USD 6,402. That cut generated 63 applications at roughly USD 101.61 CPL and −70% ROMI. Excluding the non-contributing hours stopped a loss that remained fragmented across vendor reports.

The next constraint was the page. The nutrition offer and lead magnet were revised. The published report attributes a 1.6 percentage-point increase in application conversion and a rise in sales conversion from 6% to 10% to those changes. This was not cosmetic CRO; it clarified the decision before the lead reached sales.

• Read hour, source and campaign together with revenue
• Stop spend that never reached a commercial signal
• Align offer and lead magnet with actual intent
• Judge the page by sales, not forms alone

The most valuable saving was not a cheaper click. It was no longer funding segments unable to return value.

June closed at 300% ROMI and RUB 517 CPL, about USD 6.13. The application plan was missed, yet return doubled compared with May at a similar level of spend. Measurement and correction proved more useful than chasing an overambitious volume target.

July tested new combinations and search-advertising vendors. CPL rose to USD 6.75 and applications fell by 361, but ROMI reached 407%. With roughly USD 1,066 less spend than the prior month, revenue grew 17%. August brought seasonal demand pressure; nevertheless CPL fell to USD 4.43 and applications increased 58%. Across the three months, the reports recorded USD 39,422 in spend, 7,066 applications and USD 162,914 in revenue at 313% ROMI.

The trajectory was not linear. During September and October, the team operated without the same direction, source tagging deteriorated and the published cut showed 24% ROMI. Attribution cleanup brought November back to 310%. December closed at 157% when volume targets began competing with return.

January and February improved as quality returned to the center: 266% and 251% ROMI. In March the business insisted on the lead plan while CPL was rising. The result was RUB 1,869,397 in spend, RUB 3,736,210 in revenue, RUB 1,070 CPL and 85% ROMI. The form target was met, but the business signal weakened.

• One leader must resolve the conflict between volume and return
• Specialists need autonomy inside a clear economic boundary
• Bureaucracy that delays tests creates an opportunity cost
• Sales feedback must return to media before budget changes

A lead plan can be achieved while the business gets worse. The primary metric must match the real constraint.

The first publication documented a partial checkpoint: RUB 31,153,000 in revenue, RUB 9,181,000 in spend and 239% ROMI. At the current reference rate, that is approximately USD 369,628 in revenue and USD 108,932 in spend. It should not be added to the later total.

The 2024 update extended the full period and reported RUB 103,703,385 in revenue, RUB 24,217,853 in spend, 85,275 applications and 328% ROMI. At the reference rate, those totals equal USD 1,230,433 in revenue and USD 287,343 in spend. The average CPL derived from the totals is about USD 3.37. These are two time snapshots of the same system.

Time context matters: the first total is a checkpoint, while the second replaces it as the updated cumulative result.

The historical project was multichannel: search advertising, paid social, a classified marketplace and other sources. This reconstruction removes local platform brands because the lesson does not depend on one vendor. Every source has to connect to the same economics, and budget has to follow contribution.

Today that loop often starts with Meta Ads and continues through a landing page or form, WhatsApp, CRM and sales stages. Pixel, CAPI and UTMs preserve useful signals, but do not replace commercial truth. Creative, audiences and pages are judged by quality, conversations, opportunities, customers and revenue.

• Define ticket, margin, lead quality and allowable CAC before scaling
• Connect campaigns to CRM and post-lead outcomes
• Separate production budget from learning budget
• Use cohorts when the sales cycle crosses month boundaries
• Scale in increments the operating team can absorb