The night shift

Most performance engagements begin by installing tracking. This one could not. The website, the booking system and the domain belong to the franchisor rather than to our client, so as the franchisee’s marketing partner we could not place a pixel, fire conversion events, or build a retargeting audience from site visitors.

No pixel means no website custom audiences, no lookalikes built from real customers, no retargeting of people who browsed and left, and no server-side confirmation that a click became a booking. Effectively every optimisation tool modern paid social depends on was unavailable to us.

Rather than stalling the launch until access was negotiated with the franchisor, we started without the standard setup and treated the studios themselves as the measurement layer. That only works if everyone understands their part, so we set out plainly what we were running, what signal we could and could not see, and what the studio teams needed to do at their end. The first campaigns generated qualified lead flow without any of the tracking infrastructure normally treated as a precondition.

The client then narrowed the goal from leads in general to phone-call leads exclusively, and paused paid social. We rebuilt the strategy around search accordingly.

After several months, cost per phone-call lead was running very high. Then the platform paused call-only ads, and within the remaining formats we could not remove the directions action or run a clean call-only objective. The campaign could no longer be pointed at the one outcome the client wanted.

This is a platform-capability failure rather than a media-buying one. No amount of bid, budget or creative work fixes an objective the platform will not let you isolate. We raised it immediately — before a review meeting, before a quarterly report, before the client noticed the numbers themselves — and we arrived with the replacement already designed rather than with a problem for the client to solve.

Five separate campaigns, one per location, each pointed at its own studio’s phone line. Every studio has its own catchment, its own competitive pressure and its own cost per call, so each is funded and judged on its own numbers rather than averaged into a group total. A single pooled campaign optimises toward whichever studio converts most cheaply and quietly starves the rest — which for a franchise operator whose revenue is per studio means one location subsidising the group average while its own phone stays quiet.

With no pixel and no website audiences available, targeting had to be reasoned out from who the customer actually is rather than inherited from site behaviour. Each campaign was aimed at working women inside the realistic travel radius of that specific studio, and creative was built to produce a phone call rather than a click — leading with the studio, the service and a reason to ring now.

Every studio in the group opens at 9am and closes at 7pm. The default instinct is to run ads during business hours so calls arrive when someone can answer them. We did the opposite, deliberately.

The audience for lash extensions in these catchments is overwhelmingly working women. Someone in an office from nine to five does not research a beauty appointment at 11am and does not phone a studio from her desk. She looks in the evening, at night, and early in the morning before the day starts. If that is true, running ads only during studio hours means advertising to the audience precisely when it is least able to respond — and paying peak auction prices for the privilege.

There was a platform reason pointing the same way. Restricting delivery windows on paid social repeatedly pushes campaigns back into the learning phase, so scheduling costs efficiency across the whole campaign in exchange for tidier timing. You pay more per lead to receive fewer of them. We kept delivery continuous and solved the timing problem on the business side instead.

After-hours calls cost $11.68 against $40.44 during the working day — a seventh of the budget producing two-fifths of the calls. Then we looked at what happened to those calls.

During studio hours, 21.1% of calls became a conversation lasting a minute or more. After hours, 4.1% did. Same ads, same audience, same intent — five times less likely to be answered, for the simple reason that the studio is closed and nobody is at the desk. Of 145 after-hours calls, 139 never became a 60-second conversation. At the daytime rate roughly 31 would have. And because the platform stores no phone numbers, not one can be called back.

So our best discovery was also creating the group’s biggest leak. Reporting the first half alone would have been a more flattering case study and a worse piece of work.