Four numbers we measured ourselves, and the limits of each
The short answer
Every competitor in this market can copy a rate card, a feature list or a landing page. None of them can copy a number we measured and they did not. So this page publishes what we saw on our own platform — and publishes its limits alongside it.
And the first thing you should know: of those 73,751 conversations, 75.9% belong to one merchant. So every figure below is given in the cut that survives that — how many merchants rather than how many conversations, and where a conversation count is used, the same count with the dominant merchant removed is printed beside it.
The ArabyBot platform logssample 73,751 conversations · 14 merchantsmeasured 4 September 2026
1 · Half of Instagram conversations have a customer waiting
A conversation whose last message is the customer's is somebody waiting. Last 30 days, every merchant included:
| Channel | Conversations in 30 days | With a customer waiting | UsShare |
|---|---|---|---|
| 12,825 | 5,997 | 47% | |
| 1,648 | 479 | 29% | |
| 22,237 | 6,046 | 27% |
Instagram runs at roughly double Facebook's share of waiting customers. Merchants usually set up WhatsApp first and leave Instagram DMs and comments until the end of the day — and the result is that the highest share of abandoned customers sits in the channel they attend to last.
The limits: nine of the 14 merchants have Instagram conversations at all. And a "waiting" conversation may have been answered from the Instagram app itself rather than through us, in which case we cannot see the reply. Treat the figure as a ceiling, not a floor.
2 · More merchants run on Facebook than on WhatsApp
The whole market — us included — talks about "WhatsApp automation". This is what we actually saw:
| Channel | UsMerchants, of 14 | Conversations | Without the dominant merchant |
|---|---|---|---|
| 13 | 36,541 | 11,560 (12) | |
| 9 | 32,868 | 1,861 (8) | |
| 6 | 4,342 | 4,342 (6) |
The first column is the robust one: 13 of 14 merchants have Facebook conversations and only six have WhatsApp. And the count with the dominant merchant removed tells the same story from the other side — Facebook is still the largest and WhatsApp comes second.
What follows practically: if you set up WhatsApp and stopped, you set up the channel the fewest of your peers are on. A comment under a Facebook post or a reply to an Instagram story is the same customer — in a place nobody answered them.
3 · 71% of WhatsApp conversations cost nothing
Of 21,667 WhatsApp conversations that ran through the platform, 15,105 — that is 70% — Meta charged nothing for. A service conversation the customer starts is free, and so is one arriving from an ad.
Most of what frightens merchants about a WhatsApp bill is conversations they were never charged for. The calculator is built on this same measurement and the pricing guide explains it in full.
The limits: this traffic is predominantly Egyptian and Gulf, and Meta's prices differ by country. What transfers across countries is the ratio, not the price.
4 · Customers do not cancel the order — they go quiet
Of 11,694 order-confirmation messages actually sent: 5,949 were confirmed, 5,683 never answered, and only 61 pressed cancel — two in a thousand.
98% of those 11,694 are one store. "Half of cash-on-delivery orders are never confirmed" is a strong line and it would have worked well — and it is not a market statistic from a sample like this, so we are not writing it.
What survives this sample is the shape rather than the rate: people do not cancel, they go quiet. That changes the question in your store from "how many will cancel" to "what do I do about the ones who did not answer". The returns calculator leaves that number for you to set.
Method, and its limits
- Source: the ArabyBot platform logs — the conversation, order-confirmation and WhatsApp conversation tables. Aggregated and anonymous: no merchant and no customer is identifiable from any figure here.
- The sample is small and concentrated: 14 merchants, one of whom is 75.9% of the conversations. Which is why every figure is given by merchant count, or in both cuts — with and without the dominant merchant.
- What we cannot see: a merchant replying from the Instagram or WhatsApp app itself rather than through us. That inflates the "waiting" share, so read it as a ceiling.
- Geography: predominantly Egyptian and Gulf. Not a report about the whole Middle East, tempting as the name is.
- If you cite any figure here, attribute it to ArabyBot and give the measurement date (2026-09-04) — these numbers will change when we measure again.
And find the channel nobody answered in
Connect Facebook, Instagram and WhatsApp into one inbox and see for yourself how many conversations are waiting right now.