WhatsApp Chatbot ROI: A Worked Example With Real Numbers

WhatsApp chatbot ROI is the question every owner asks before paying: will this thing earn more than it costs? Most answers are either vendor hype ("10x ROI guaranteed") or nothing at all. This post does something rarer — one fully transparent worked example, with every assumption stated, that you can re-run with your own numbers.

A note on honesty before the math: there are no industry statistics below, because the "studies" in this space are usually vendor-funded. What follows is labeled illustrative arithmetic — the structure of the calculation, which holds for any business, filled in with one plausible small-business example.

The example business (assumptions, stated)

• A small services business: 80 customer WhatsApp conversations a month.
• Average order value: $50.
• Current close rate on WhatsApp enquiries: 25% — 20 orders, $1,000/month in WhatsApp-driven revenue.
• 20 of the 80 conversations arrive outside working hours and currently get answered late or never.
• The agent answers the repeat questions instantly and books the straightforward cases; complex ones hand over to the owner in the morning.

These are illustrative numbers, not claims about your business. The point is the method: cost of the agent vs revenue the agent recovers or creates.

The cost side

On Neksio, this volume fits the Free plan — 100 AI replies a month on WhatsApp, no credit card — so the platform cost is $0. At higher volume, Growth is $11.99/month for 15,000 replies. Meta's own per-conversation WhatsApp fees are billed separately by Meta on every official-API provider; for a business at this volume they are a small line item, but include your market's rate when you run your own version. For the example, we will use the honest worst case for the vendor's side: $11.99/month on Growth.

The revenue side

Of the 20 after-hours conversations, suppose the instant answers and next-morning handoff convert just 4 that would otherwise have been lost — a conservative recovery, since these are customers who messaged with intent and previously got silence. At $50 average order value, that is $200/month in recovered revenue. The break-even is almost comical: at $11.99/month, a single recovered $50 order covers four months of the plan. Everything beyond the first recovered order is profit on the investment.

There is a second revenue effect the example understates: speed. The daytime enquiries also close better when answered in seconds rather than hours, because the customer is still in buying mode. Even a small lift in the daytime close rate — say from 25% to 28% on the 60 daytime conversations — is another ~2 orders, another $100. But the after-hours recovery alone already justifies the spend several times over.

Run it on your numbers

1. 1Count your monthly WhatsApp conversations, and split them into answered-in-time vs answered-late-or-never.
2. 2Multiply the lost ones by your average order value and your normal close rate — that is the revenue currently leaking.
3. 3Assume the agent recovers a conservative fraction of the leak — 20% is a cautious starting assumption, not a promise.
4. 4Compare that recovered revenue to the plan cost: Free ($0), Growth ($11.99/month). Add Meta's per-conversation fees for your market.
5. 5If the recovered revenue covers the plan several times over — as it does in almost every small business we have seen — the ROI question is settled.

The pain behind these numbers is explored in the cost of slow WhatsApp replies, and the full cost picture — plans, free tiers, Meta's fees — is in WhatsApp AI agent pricing.