Every founder we talk to has been quoted three completely different numbers for WhatsApp marketing, and none of the three arrived with a line-item breakdown. One vendor says fifteen hundred rupees a month. Another says two lakh. A third says it depends, then sends a proposal with a single figure buried on page nine.
The spread is real, but it is not arbitrary. WhatsApp marketing in India has three separate cost layers that get bundled and quoted as one price: the software you send from, the conversation fees you pay regardless of software, and the human or AI labour that actually answers the customer. Vendors blur those layers because blurring them is where the margin lives.
We run WhatsApp at volume on our own brands before we run it for anyone else — 24/7 AI sales agents on every number we own, a 33 second average reply time, roughly 80% of inbound support resolved before a human opens the inbox. So this is the breakdown we would want if we were the ones buying. Every figure below is either a market rate you can verify yourself or a number out of our own stack.
The three layers you are actually paying for
Before you compare any two quotes, separate them into layers. A quote that does not separate them is not a quote, it is a hope.
Layer one is the sending platform — the software that connects to the WhatsApp Business API, holds your templates, runs your broadcasts and gives your team an inbox. Layer two is conversation pricing, metered per conversation window and billed through whichever platform you sit on top of. Layer three is answering capacity: the people or the AI agents who handle what comes back. Layer three is the one everybody underprices, and it is the one that decides whether the other two were worth paying for.
The trap is structural. Layer one is cheap and visible, layer two is moderate and metered, layer three is expensive and invisible until launch week. Brands sign up for a cheap platform, broadcast to forty thousand contacts, get two thousand replies in an evening, and discover they have a two-person team and no answer for any of them. The software cost was never the constraint.
- Layer 1 — Platform: self-serve tools run roughly $32–82 a month at the tiers most consumer brands land on.
- Layer 2 — Conversation fees: metered per conversation window, unavoidable on any platform.
- Layer 3 — Answering capacity: humans on shift, or AI agents. This is the layer that breaks first.
- Bundled retainers hide all three inside one number, usually with layer three quietly capped.
What the self-serve tools really cost
The platform layer has commoditised hard. A functional WhatsApp Business API tool with templates, broadcasts, a shared inbox and basic automation sits in the $32–82 a month band for the plans most consumer brands actually use. That is genuinely cheap, and if all you need is a broadcast button and a place for two people to reply from, buy the cheap tool and stop reading.
The reason most brands do not stop there is that the entry tier is priced for low conversation volume and simple flows. The moment you want segmentation that reflects purchase history, journeys that branch on what someone actually replied, multiple numbers with shared context across city or category teams, or an agent that can quote your real pricing logic rather than push a menu tree, you are either climbing tiers fast or rebuilding the logic yourself in a workflow tool alongside it.
The other thing the platform tier does not include is setup. Business API access, number verification, display-name approval, template approval, and the green tick where you qualify — each is a small piece of admin that eats a fortnight if nobody on your team has done it before. It is not expensive work. It is work that has to be done correctly once, and it stalls launches constantly.
Where managed retainers land, and why
Managed WhatsApp sits in a completely different price band, and it should — you are buying layer three. Our WhatsApp Engine starts at $1,500 a month as a standalone system: Business API setup and verification, an AI sales agent trained on your catalogue, pricing rules and objections, a multi-number shared inbox with routing and assignment, broadcast templates and segmentation, and the always-on journeys — welcome, abandoned cart, post-purchase, win-back, re-order.
Inside a broader engagement it stops being a line item. Our Growth tier is $7,500 a month and bundles WhatsApp journeys with the Content Engine, Social Engine and Outreach Engine, because a WhatsApp number with nothing driving traffic to it is an expensive answering machine. The full stack — all eight systems, Voice included — is $15,000 a month at the Dominator tier.
Compare that against the honest cost of layer three staffed by humans. Twenty-four hour coverage on a single number is not one salary, it is three shifts plus holiday cover plus attrition plus the retraining cycle every time somebody leaves. And it still will not average a 33 second reply at eleven at night, because nobody does.
- Self-serve platform only: roughly $32–82 a month, plus conversation fees, plus your own people answering.
- WhatsApp Engine, standalone: from $1,500 a month — setup, AI agent, multi-number inbox, journeys, broadcasts.
- Growth tier: $7,500 a month — WhatsApp journeys with content, social and outreach feeding the same funnel.
- Dominator tier: $15,000 a month — all eight systems including Voice Engine campaigns.
- Human 24/7 coverage: three shifts minimum, and still no 33 second average after midnight.
The numbers that decide whether it pays back
Cost only means something against throughput, so here are the operating numbers from our own stack rather than a cherry-picked case study. A 33 second average reply time, day or night. Roughly 80% of inbound support resolved without a human opening the thread. Agents live 24/7 across every number on the account, with one inbox holding shared context so a conversation keeps its history when it moves between an AI agent and a person.
Two of those matter more than they look. The 33 seconds is not a vanity metric — in India the sale happens on WhatsApp and it happens in the first minute, so reply latency is conversion, not customer service. The 80% automation figure is the one that changes your headcount plan: your existing team handles the fifth of conversations that need judgement instead of drowning in the four fifths that do not.
The honest caveat is that neither number appears in week one. The agent goes live supervised, with a human watching every thread for the first fortnight while corrections feed straight back into its instructions. Accuracy climbs fast in that window. Anyone selling full autonomy from day one is selling a bot that will confidently quote a wrong price, and a confidently wrong price costs more than a slow reply.
There is a second-order effect worth pricing in too. Because every reply is instant, the funnel above WhatsApp gets more valuable without changing spend — the same ad, the same sampling drop, the same screen impression converts harder when the follow-up conversation starts in half a minute rather than the next morning.
How to price this for your own brand
Start from conversation volume, not from a plan comparison page. Estimate how many inbound conversations a month you would have if every enquiry, every abandoned cart and every post-purchase question came through WhatsApp rather than scattering across email, DMs and phone calls. That number, times your conversation rate, is your unavoidable floor. Everything else is a choice about who answers.
Then decide honestly whether layer three exists inside your company. If you have a support team with spare evening capacity and a catalogue simple enough to answer from memory, a self-serve tool at $32–82 a month is the right call and you should not pay anyone a retainer. If your answer volume already exceeds what your team can hold, or the questions require pricing logic, availability checks and objection handling, you are choosing between hiring shifts and running agents.
The build-versus-buy line usually falls where setup complexity meets always-on coverage. Buying the software is easy. Buying the software, wiring the journeys, training an agent on your actual catalogue, keeping it accurate as offers change, and having it answer at 11pm on a Sunday is the part with a real price on it. Whichever way you go, insist the quote separates the three layers — a vendor who will not itemise them is telling you something.
Questions we get asked
How much does WhatsApp marketing cost per month in India?
It splits three ways. Self-serve platform tools land roughly in the $32–82 a month band for typical consumer-brand plans, conversation fees are metered on top regardless of platform, and then you pay for whoever answers. Fully managed sits far higher because it includes that answering layer — our WhatsApp Engine starts at $1,500 a month with a 24/7 AI sales agent, multi-number inbox and journeys included, and $7,500 a month at Growth tier where content, social and outreach feed the same funnel.
Is a WhatsApp marketing agency worth it versus doing it in-house?
In-house wins when your conversation volume fits inside your existing team's capacity and your catalogue is simple enough to answer from memory. It stops winning the moment you need 24-hour coverage, because that is three shifts rather than one hire, and it still will not produce a 33 second average reply after midnight. The real comparison is not agency versus in-house, it is what the answering layer costs either way.
How fast does a WhatsApp reply need to be to convert?
Inside the first minute. In India the purchase decision on WhatsApp is largely settled in the opening exchange, which is why we hold a 33 second average across our own numbers day and night. A reply that lands the next morning is competing against whoever answered last night, and it usually loses regardless of how good the answer is.
Systems behind this playbook
Want this priced against your actual conversation volume? Bring us the numbers and we will show you the maths before anyone signs anything.
Book a demo