Trumpet Lessons Case Study

Coursdetrompette.fr is the business of a trumpet teacher with more than twenty years of teaching behind him, online and in students' homes near Lyon, for pupils aged 7 to 87. The site ranks well on Google, but "trumpet lessons" in French is a narrow market: fewer than 200 visits a month.

The brief was simple: find out whether Facebook could bring in students, on a total budget of €100 at most, approved before launch. We said from day one that at this budget, one measure would really count: the number of contacts gained.

The ad shows what Emilien can play after six months of lessons. A short, sincere text with no marketing promise, for an audience that often hesitates to start.

It showed from the second day: 214 people watched the whole video in a single day, half of all viewers stayed to the end, and the ad was being shared. Even people who were not looking for lessons stopped to see how far Emilien had come.

Contact goes through a pre-filled Meta lead form: two taps, without leaving Facebook. Messenger stays open for those who would rather write to the teacher.

Targeting aimed at men aged 25 to 45, with the only relevant interest available: musical instruments. Instagram was removed, as the video without text was too weak there. Meta's Audience Network was kept for its very low cost per thousand impressions.

The ad was approved within minutes on May 20, then Meta took its time analysing news feeds before spending its first cent. After a week the campaign was out of its learning phase: 10 leads for €22, or €2.26 per lead against a €3.08 median for comparable advertisers.

On May 24, Meta reported an unexpected audience: mostly people over 65, and more women than men. We changed nothing. It was too early, and nobody yet knew which of these contacts would become students. On a budget this size, an A/B test would mostly have risked breaking what worked.

On June 8 the campaign stood at 23 leads for €66, with a cost per lead 14% below similar ads. On June 11, 30 leads in three weeks. Halfway through, we forecast around forty for the month.

The campaign ended on June 19 with 45 leads. As the client put it on June 8: "I think we now have enough perspective to say this is no longer just luck."

With the test validated, the next campaign moved to €12 a day, on Meta's estimate of about five leads a day at that budget. The two campaigns overlapped for a day to avoid any gap. A payment incident on Meta's side stretched the overlap, which brought in 14 leads in two days.

The new campaign, however, missed its learning phase. We paused it, then relaunched it with a cost per result goal capped at €3. Cost first dropped below €4, then climbed back towards €6: on a small daily budget, a cap is only a guide. The campaign stopped as planned on July 8, with its average cost back to normal.

A teacher knows better than anyone when his slots are full, when the holidays start and when term begins again. The next campaign therefore runs on a daily budget, capped at €3 per lead, and the client turns it on or off to suit his diary.

To avoid ad fatigue, the logical next step is to film other students at the same six-month stage: more proof, and an ad that keeps renewing itself.

By June 11 the teacher had received 8 lesson requests and answered each one, by Messenger, WhatsApp or text message. A contact goes from warm to cold within 48 hours: speed of reply matters as much as cost per lead.

Contacts now land in a CRM the client owns, on his own hosting, with a guide and a video course sent automatically. Next step: send the contacts who became students back to Meta through the Conversions API, which Meta estimates could lower the cost per qualified lead by 24%.