A Sky Designers performance case study — SITC Campus HRM Diploma, across Meta, TikTok and Google Ads
LKR 936,143 of advertising across three channels produced 972 confirmed registrations worth LKR 23,328,000 — a blended return of 24.9x, at a cost per registration of LKR 963.
Advertising came to 4.0% of booked revenue.
But the most useful number in this report isn't the 24.9x. It's what happens when you put this campaign next to a near-identical one the same account ran for a different course.
The Scorecard
| Metric | Result |
|---|---|
| Total ad spend | LKR 936,143 |
| Impressions (all channels) | 4.64M |
| WhatsApp conversations | 5,318 |
| Confirmed registrations | 972 |
| Booked revenue | LKR 23,328,000 |
| Blended ROAS | 24.9x |
| Cost per registration | LKR 963 |
| Ad spend as % of revenue | 4.0% |
Meta carried the budget — nine rupees in every ten. TikTok took 9.2% and Google Ads ran at test scale, under 1%. On Meta spend alone, the return was 27.6x at LKR 868 per registration.
The Comparison That Matters
The same campus ran an English Diploma campaign through the same tracker, the same team, the same WhatsApp funnel. Here is how the two compare on Meta spend alone:
| Metric | HRM Diploma | English Diploma |
|---|---|---|
| Course price | LKR 24,000 | LKR 13,000 |
| Registrations | 972 | 1,628 |
| Cost per registration | LKR 868 | LKR 1,430 |
| Conversation → registration | 18.3% | 14.9% |
| Meta spend as % of revenue | 3.6% | 11.0% |
| Return on Meta spend | 27.6x | 9.09x |
The higher-priced course returned three times more per rupee — and it did so while costing less to acquire each student, not more.
That last part is counter-intuitive and worth sitting with. The instinct is that a more expensive product is harder to sell, so acquisition should cost more. Here the opposite happened: the LKR 24,000 course converted chats to registrations at 18.3% against 14.9% for the LKR 13,000 course.
Why Price Point Beats Optimisation
There is a ceiling on what campaign optimisation can achieve. You can cut cost per conversation, improve creative, tighten the funnel — and all of that is real work that moves real numbers.
But none of it moves the needle like what you're selling.
A course at LKR 24,000 generates 1.85x the revenue per registration of one at LKR 13,000. That multiplier applies to every single sale, automatically, with no optimisation effort at all. It is the single largest lever in the entire account, and it sits outside Ads Manager entirely.
This is why the question "how do I improve my ROAS?" often has a commercial answer rather than a media one. Before rebuilding campaigns, it's worth asking whether the highest-value thing you sell is the thing you're advertising.
Concentration, Again
All LKR 844,083 of Meta spend ran through two campaigns. Not eighteen. Two.
The English Diploma campaign ran 18 campaigns to produce 9.09x. The HRM campaign ran two and produced 27.6x.
Fewer campaigns means each one accumulates more conversion signal, exits the learning phase faster, and gets optimised harder. Splitting the same budget across eighteen structures gives the algorithm eighteen small, slow-learning datasets instead of two strong ones.
Where TikTok Fits
TikTok took 9.2% of the budget and delivered impressions at LKR 120 CPM against Meta's LKR 217 — roughly half the cost. Clicks came in at LKR 1.24 against Meta's LKR 64.
Video also held far better there. TikTok recorded a 30.9% two-second view rate, and six in ten of those viewers continued to six seconds. On Meta, hook rate for the same creative approach was 0.45%.
The honest caveat: TikTok's recorded conversions are platform events, not course registrations, and there's no conversation-level data linking TikTok to an enquiry. What the numbers support is that TikTok bought attention cheaply. What they don't yet support is a claim about registrations. Getting that answer requires conversation-level tracking on the TikTok side — which is the clearest next step for this account.
A Note on Honest Measurement
The 972 registrations are a single course total. They are not split by channel, and the three platforms report on different windows and different definitions — Meta reports link clicks, TikTok reports destination clicks, Google reports clicks, and none of the three CTR figures are measured the same way.
It would be easy to present the blended 24.9x as though all three channels contributed proportionally. They may have. The data doesn't prove it either way.
Reporting that overstates certainty is worse than reporting that admits its limits, because the next set of budget decisions gets made on it.
The Takeaway
Two campaigns, one channel carrying 90% of the budget, a straightforward WhatsApp funnel — and a 24.9x blended return.
The campaign structure was simple. The thing that made the numbers exceptional was the price of what was being sold.
Optimisation improves a campaign. Selling something worth more changes the arithmetic.
Sky Designers is a Colombo-based digital marketing agency specialising in performance advertising across Meta, TikTok, and Google. Want to know which of your products your advertising should actually be behind? Get in touch.