A Sky Designers performance case study — Rysera, AI Course and Robotic Course, Meta ads, 1 August to 10 November 2026
Rysera spent LKR 1,404,241 on Meta ads for its AI and Robotic courses over just over three months, collected LKR 6,631,565 from course buyers, and returned 4.72× on ad spend.
But the headline number isn't the useful part. Two courses at very different prices share one budget and one cost per buyer, and that single figure fits neither of them. The way Meta's enquiries line up against paid enrolments also needs a closer read. Both points apply to any Sri Lankan course provider selling more than one programme.
The Numbers
| Metric | 1 Aug to 10 Nov 2026 |
|---|---|
| Ad spend | LKR 1,404,241 |
| Message enquiries | 4,836 |
| Cost per enquiry | LKR 290 |
| Course buyers | 768 |
| Buyers per 100 enquiries | 15.9 |
| Ad spend per buyer | LKR 1,828 |
| Paid revenue | LKR 6,631,565 |
| Return on ad spend | 4.72× |
Revenue is the amount buyers actually paid, counted to 9 November, not the list price. Spend is from Meta Ads Manager. Buyers and revenue aren't split by ad set.
| Course | Buyers | Average paid (LKR) | Revenue (LKR) | Share of revenue |
|---|---|---|---|---|
| AI Course | 668 | 5,941 | 3,968,565 | 59.8% |
| Robotic Course, online and physical | 100 | 26,630 | 2,663,000 | 40.2% |
| Total | 768 | 8,635 | 6,631,565 | 100% |
Why one cost per buyer can't fit two courses
The Robotic Course had 13.0% of the buyers and brought 40.2% of the revenue. Each Robotic buyer paid about 4.5 times what an AI Course buyer paid on average.
Set the blended LKR 1,828 of ad spend per buyer against each course and the picture splits. It's about 30.8% of an average AI Course payment, and about 6.9% of a Robotic Course payment. One looks like a tight margin and the other looks like room to spend far more. Neither is the real figure, because buyers weren't traced to ad sets, so we can't say what each course actually cost to sell.
If you sell more than one course, give each its own ceiling: the most you can pay in ads for one buyer of that course. Then run each course in its own campaign, or at least tag every buyer with the course and the ad that started the chat. A higher-priced course may deserve a much larger share of the budget than a blended figure suggests.
Did engagement ads really beat sales ads?
| Ad set type | Spend (LKR) | Share of spend | Message enquiries | Cost per enquiry (LKR) |
|---|---|---|---|---|
| Engagement ads | 516,141 | 36.8% | 2,447 | 211 |
| General audience | 269,510 | 19.2% | 973 | 277 |
| Sales ads | 179,686 | 12.8% | 849 | 212 |
| Video ads | 160,451 | 11.4% | 532 | 302 |
| Custom audience | 18,117 | 1.3% | 35 | 518 |
| Awareness, website traffic, reach and post ads | 260,336 | 18.5% | n/a | n/a |
| Total | 1,404,241 | 100% | 4,836 | 290 |
They tied. Engagement ads brought enquiries at LKR 211 and sales ads at LKR 212. Together they took 49.6% of spend and brought 68.2% of enquiries. General audience and video ads took 30.6% of spend at LKR 277 to 302.
So the name of the objective mattered less than you'd expect. What these costs can't show is which enquiries became buyers. An engagement ad can bring plenty of chats from people who are curious but not ready to pay. Until each buyer is linked to the ad set that brought them, treat the LKR 211 as a cost per chat, and judge any shift in budget on buyers.
What should 18.5% on views and visits have to prove?
Awareness, website traffic, reach and post ads took LKR 260,336 and weren't set up to bring messages. The awareness video ads brought 423,820 full video views at about LKR 0.47 each. The website traffic ads took LKR 52,296, and this report doesn't count the visits or the enrolments that came from them.
That matters because revenue counts every paid enrolment, whatever route the buyer took. Some of the 768 may have enrolled on the website or after watching a video without ever sending a message. Count website enrolments separately, and ask each new student where they first saw the course. Then you'll know whether this 18.5% is building demand or just buying cheap views.
Where the cheapest growth sits: 15.9 buyers per 100 enquiries
768 buyers against 4,836 enquiries is 15.9 per 100. It isn't a true close rate, since some buyers may never have sent a message, and some enquiries may have come from people who enrolled earlier. It is still the best signal in this report of how well chats turn into paid students.
As an illustration, one more buyer for every 100 enquiries is about 48 more buyers. At the average of LKR 8,635 paid, that's about LKR 417,600, taking the return from 4.72× to about 5.02× with no extra ad spend. Record where each chat stops: at the fee, at the schedule, at the payment details, or with no reply. Our guide to tracking enquiry-to-booking rate shows how we set that up.
The Takeaway
When two courses share a budget, a single cost per buyer hides what each one can afford. Give every course its own ceiling, judge ad sets on buyers rather than chats, and count website enrolments so the spend on views and visits can be judged too. When you plan the next intake, our method for setting a budget from a sales target works from these same rates. More on how we run these campaigns on our performance advertising page.
Sky Designers is a Colombo-based digital marketing agency specialising in performance advertising across Meta, TikTok, and Google. Selling more than one course and not sure which one your ads should push? Get in touch.