- benchmarks
- data
- South Africa
Every brief that lands on a marketing director’s desk asks the same question. What should we expect? And the public answers are usually scarce, dated, or shaped by whoever wants to sell you something. This is the practitioner-honest version of the working ranges we use to forecast SA mobile campaigns in 2026.
How to read these numbers
These are working ranges, not a forecast. The variance inside each range is wide. The right number for any specific brief depends on the mechanic, the audience profile, the prize structure, the supporting media spend and the time of year. Treat what follows as the starting input to a campaign model, not the output.
If a vendor or agency quotes you a “typical” number with no range and no caveats, ask what data the typical is averaged across and what the standard deviation is. Mobile marketing benchmarks are too campaign-specific to honestly cite as a single number.
Channel-level performance ranges
USSD promotional campaigns
USSD entry funnels in the SA market typically perform in these ranges for a well-designed three-screen menu.
| Metric | Working range | What moves it |
|---|---|---|
| Entry-screen reach | 65 to 80% of dialled-in consumers | Menu length, clarity of first screen, prize attractiveness |
| Completion (entry submitted) | 50 to 72% of those reaching entry screen | Number of menu steps, validation difficulty, session timeout |
| Cost per entry (free-to-consumer) | R0.18 to R0.45 | Network mix, contracted volume tier |
| Daily entry volume on flighted campaign | 800 to 12,000 entries per day | Media spend, prize tier, on-pack distribution |
| Repeat-entry rate | 1.4 to 3.2 entries per unique entrant | Mechanic (one-per-day vs unlimited) |
The 50 to 72% completion range is what surprises clients most. People assume USSD is a clean funnel because the channel is simple. It is not. Every screen sheds entrants. A poorly designed five-screen menu can lose 40% by screen three. The 72% end of the range belongs to menus with three screens or fewer.
WhatsApp Business campaigns
WhatsApp engagement runs higher than email or SMS on the open and reply numbers. The ranges below assume a Meta partner setup and properly approved template messages.
| Metric | Working range | What moves it |
|---|---|---|
| Opt-in rate from landing page | 18 to 34% | Incentive strength, click-to-WhatsApp friction |
| Template message open rate | 75 to 92% | Template category, time of send |
| Reply rate to active conversation | 38 to 65% | First-message hook, time-of-day |
| Cost per conversation (SA marketing category) | R0.50 to R1.05 | Volume tier, brand account configuration |
| 30-day re-engagement rate | 12 to 28% | Initial mechanic, follow-up content relevance |
Open rates above 90% sound implausible. They are not. WhatsApp open rates are genuinely that high because the inbox is uncluttered relative to email and the channel is conversational by default. The number to watch is the reply rate, which is where the funnel starts doing real work.
Premium SMS, for reference
Premium SMS is in slow decline but the benchmarks still matter because some clients ask for SMS-versus-WhatsApp comparisons.
| Metric | Working range |
|---|---|
| Cost to consumer | R1.50 (CPA Reg 11 cap) |
| Cost to brand (free-to-consumer alt) | R0.25 to R0.45 per entry |
| Delivery rate | 95 to 99% |
| Response rate to two-way SMS prompt | 6 to 18% |
The response-rate number is why SMS has lost share to WhatsApp and USSD for any two-way mechanic. It works for broadcast. It does not work for competitions or surveys.
Audience patterns
Network mix in a typical SA consumer campaign
National-flighted FMCG promotions tend to break out across the four networks in something like the proportions below.
- Vodacom: 40 to 52% of entries
- MTN: 26 to 38%
- Cell C: 6 to 14%
- Telkom Mobile: 4 to 9%
Two things to read into this. If a campaign is showing Vodacom share above 60% or below 30%, something is unusual. Either the media plan is skewed (radio buy concentrated in Vodacom-stronger regions), the prize is over-indexed to a specific demographic, or there is a technical issue on one of the other networks. Investigate before the campaign closes.
The other read is that the bottom-tier numbers still matter. A 4% Telkom Mobile share is still 4% of the audience. The WASPA Code requires the campaign to work on every network the brand has not explicitly excluded.
Time-of-day patterns
USSD entry volume on a typical FMCG promotion peaks in two windows. The first is 11:00 to 13:00, and on Fridays the lunchtime spike is sharper than the rest of the week. The second is 18:00 to 21:00 every day.
WhatsApp engagement has a different shape. Peak is 19:00 to 22:00, post-dinner couch time. Strong weekend daytime engagement happens that USSD does not see, because the smartphone audience is on Wi-Fi on weekends.
Brands sending WhatsApp broadcasts at 10am on a Monday and wondering why engagement is low. There is the reason.
Day-of-week patterns
Saturday and Sunday usually pull 1.4 to 2.1x the weekday entry volume on consumer FMCG promotions. The exception is when on-pack distribution is heavily skewed to forecourt convenience stores, where the Friday afternoon spike is sharper than the weekend.
Want us to model your specific brief?
If you have a campaign in planning and want a forecast that uses these ranges with your actual media plan and prize budget, send us the brief and we will model it.
Mechanic patterns worth knowing
A few mechanic-level patterns hold across most consumer promotional work.
Pin-code on-pack mechanics. Entry-to-redemption rates land in the 8 to 22% range. Variance is huge and driven mostly by prize attractiveness. Cash prizes pull harder than vouchers. Vehicle giveaways pull hardest.
Till-slip mechanics. Lower volume than pin-codes because of the verification step, but higher quality entrants because you know they actually purchased. Typical entry volumes run 30 to 55% lower than the equivalent pin-code mechanic.
Quiz-based mechanics. A single yes-no question increases entries by around 12%. A three-question knowledge quiz drops entries by 35 to 50% but lifts data quality and engagement time per entrant by 4 to 7x.
Loyalty-club enrolment via mobile. Completion rates of 28 to 44% for a five-field signup. The number drops to 12 to 20% if you add a sixth field, especially if that field is the ID number.
Prize tier patterns
A pattern that holds across most consumer promotions we model is the lift from a hero prize on top of the smaller prize pool.
| Hero prize value | Typical lift in entry volume vs no hero |
|---|---|
| R10,000 cash | 1.4 to 1.9x |
| R50,000 cash | 2.2 to 3.1x |
| R250,000 vehicle | 3.5 to 5.8x |
| R1 million-plus aspirational prize | 6.0 to 12.0x |
The aspirational prize tier tends to outperform purely on headline effect, even when the actual odds are worse than for the smaller prizes. A house, an overseas trip for a family of four, a year of school fees. Useful to know when sizing the prize pool against the campaign budget.
A few honest caveats about these numbers
These ranges are aggregated working numbers used inside Vaultbook to forecast SA campaigns. They are not published per-client data and we have not audited them through a third party. Your specific campaign may land outside any range here for legitimate reasons, particularly if the mechanic is unusual or the audience is niche.
A few things to keep in mind. These are South Africa numbers. They do not apply to Nigeria, Kenya, Ghana or the broader continent. The network structures, consumer behaviour and per-entry economics are different in each market. The bottom of every range comes from campaigns that were poorly designed or poorly supported by media. The top comes from campaigns where every choice was correct. Most campaigns land in the middle.
These are 2024 to 2025 working numbers and we will refresh them annually. WhatsApp pricing in particular is moving as Meta tunes the conversation-category model, and the next refresh may show meaningful drift on those rows.
How to use these benchmarks
The cleanest way to use this kind of benchmark data is in a simple two-step model.
Step one: pick the channel and mechanic. Pull the cost-per-entry range and the conversion range for that combination. Step two: multiply the expected media reach by the entry rate at the conservative and optimistic ends of the range. This produces a 3 to 5x range on forecast entry volume, which is the honest precision for a brief that has not been live yet.
Size the prize pool against that forecast volume, model the per-entry economics, then iterate. If the math does not work at the conservative end of the range, the campaign does not work. If it only works at the optimistic end, the plan is too fragile.
If you want a sanity-check on a forecast you have already built, or want us to model with you using your real media plan and prize budget, get in touch. The first conversation is always free.
Need a campaign forecast for a real brief?
If you have a campaign in planning and want a model that uses these ranges with your actual media plan, send us the brief and we will come back inside one working day.