How we helped a travel company attract customers worth 3,068,960,000 Indonesian rupiah, or 185 thousand dollars
Context and problems
The project had been advertising on Instagram for years, but measuring the effect of those ads was quite hard.
โ The main difficulty was that the sales funnel ran entirely through messenger chats: ad โ jump to WhatsApp โ deal closed in the chat. That made it hard to analyze results. It was unclear whether the ads brought customers or whether other channels deserved the credit
โ On top of that, leads from Instagram ads were very often unqualified โ either people who never replied, or completely irrelevant requests from locals
The task
Build a system for acquiring qualified leads for helicopter tours
The solution we found
- The obvious first solution was to change the funnel. Stop running ads the usual way, build a website / lead form / quiz, and then setting up end-to-end analytics becomes fairly simple.
But in the end we dropped this idea. If customers find it convenient to do it this way, we have to work with what we have.
You cannot build a business around the capabilities of technical services
The solution turned out to be fairly simple: we created several WhatsApp phone numbers and tested hypotheses on each number, recorded the results once a week and made decisions on the hypotheses.
Together with the client we introduced weekly reporting, brought all the data into one system and could clearly see how many leads, qualified leads and sales the ads produced.
- How to fight spam?
Even before our cooperation there had been attempts to solve the spam problem in many ways: stating the price in the creative, changing the language settings in the ad account, excluding Android devices from delivery, setting up interests.
We found the solution while testing yet another round of hypotheses, when the analytics we had built (it let us see not only the number of quality leads but also where the spam was coming from) showed that traffic from certain hypotheses brought many times more spam leads than other campaigns. So by solving the analytics problem we were able to understand the cause of the spam requests.
After discussing it with the client, we decided not to disclose the specific conclusions to keep the advantage.
- What made scaling possible?
Today the only way to get Instagram ads to pay back well is to work on the ad content and the sales funnel.
Settings inside the ad account can barely affect effectiveness, so 80% of our work is generating and producing various ad layouts.

Results
Budget spent: 992,072,073 Rp ($60,050)
Leads: 18,420
Cost per lead: $3.26
Qualified leads: 12,355
Cost per qualified lead: $4.86
Orders: 239
Revenue: 3,068,960,000 Rp ($185,750)
ROMI: 309%
Takeaways
-
Content is responsible for 80% of the ad result. If the ads are not working and you are told โweโll tweak the settings nowโ while the layouts stay the same โ most likely nothing will change, something has to change
-
If you do not know how to track the results of your ads and your contractor does not know either, you most likely will not get the result you expect โ look for someone who does



