How Do You Price an API Based Messaging Service for Business

The short version: price an API based messaging service by first working out your true all-in cost per message (carrier fee plus infrastructure plus support plus a failure buffer), then choose a unit customers understand, then build in margin of at least three to five times that cost before you touch a tiered plan. Most businesses skip straight to picking a number that “feels” competitive with Twilio and lose money for a year before they notice.

Why messaging pricing trips people up more than SaaS pricing

With most software, your marginal cost per extra user is close to zero. With messaging, it isn’t. Every SMS, WhatsApp message, or push notification you send has a real cost attached from the carrier or platform underneath you, and that cost changes depending on the country, the message type, and even the time of month if a carrier reroutes traffic. I’ve sat in on pricing calls where founders were confidently quoting flat per-seat pricing without knowing what a single outbound SMS to Brazil cost them. It cost more than double what it cost to send to the US. That gap eats margin fast at scale.

This is the bit that makes messaging pricing different: you’re not just selling software, you’re reselling a commodity (the message itself) wrapped in software (the API, the dashboard, the delivery logic). Get the wrapper priced wrong and the commodity part will bankrupt you quietly.

The three pricing models that get used

Strip away the marketing language and there are really three models in play.

• Per message / per conversation. You charge a fixed rate for every message sent, or for WhatsApp specifically, per conversation window (usually 24 hours). Twilio charges roughly $0.0079 per SMS segment sent within the US as of early 2026, while WhatsApp Business API pricing runs on conversation categories: authentication, utility, marketing, and service, with marketing conversations costing the most in most countries because Meta treats them as promotional traffic.
• Tiered or per-seat. You bundle a message allowance into a monthly plan, similar to how Intercom or Twilio’s own SendGrid do it. Customers pay £49, £99, £249 a month and get an allowance, with overage charged separately.
• Hybrid usage-based. A base platform fee (covers the dashboard, integrations, support) plus a per-message rate on top. This is what most serious B2B messaging vendors land on eventually because it protects your margin on infrastructure while still scaling fairly with volume.

If I’m advising a client building or reselling one of these services, I almost always push them toward the hybrid model. Pure per-message pricing punishes your heaviest, most loyal customers the most, which is backwards. Pure tiered pricing punishes light users who then churn because they feel they’re paying for messages they never send. The hybrid model is more work to build but it’s the one that survives contact with real customers.

What your real cost per message looks like

Before you can price anything, you need your true cost stack, not just the headline carrier rate. Here’s what that stack usually contains:

• Carrier or platform fee (the Twilio, Vonage, Sinch, Infobip, or Meta rate for that specific country and message type)
• Your infrastructure cost (servers, queuing, logging, webhook retries)
• Support overhead, allocated per message (this is the one almost everyone forgets)
• A failure buffer, because 3 to 8 percent of messages fail delivery on first attempt and get retried, which doubles or triples the cost of that message

So if a carrier charges you $0.0079 per SMS, your real cost once you add infrastructure, a slice of support time, and a failure buffer is closer to $0.011 to $0.014. That’s not a small rounding error at 500,000 messages a month. It’s the difference between a healthy margin and a business that quietly bleeds cash every quarter while the founder wonders why cash flow feels tight despite “good” revenue numbers.

A real story: the salon booking tool that nearly broke even at scale

Years ago, before I rebuilt my consulting work around AI, I advised a small UK software team building an appointment reminder tool for hairdressers and salons. Nothing fancy, just automated SMS confirmations and reminders sent through an API wired into a booking calendar. They priced it at £39 a month for up to 500 messages, which sounded generous and competitive against a couple of bigger UK players charging £49 to £59.

The problem: their real cost per SMS, once you added the aggregator fee, retries for failed numbers, and a support person fielding “why didn’t my client get their reminder” tickets, worked out closer to £0.03 per message rather than the £0.018 carrier rate they’d based their pricing on. At 500 messages a month, that’s £15 in real cost against £39 revenue, which looks fine on paper. But their biggest customers, chains with six or seven branches, were sending 4,000 to 6,000 messages a month on an unlimited-sounding “Pro” tier priced at £129. Real cost on that tier was closer to £150 to £180 once support time was allocated. Their best, biggest, most loyal customers were the ones losing them money every single month. That’s the uncomfortable bit nobody tells founders when they’re excited about landing a big client: growth on the wrong pricing model doesn’t fix your margin problem, it multiplies it.

We rebuilt the pricing around a hybrid model: a lower base fee, a generous but not unlimited allowance, and a clearly stated overage rate per message beyond that. Churn dropped after the change, because customers who’d been quietly worried about “unlimited” plans disappearing understood exactly what they were paying for.

Step by step: setting your actual price

1. Pull real quotes from at least three carriers or platforms for the countries your customers message into. Don’t use the US headline rate if half your customers are messaging into India, Brazil, or the Philippines, where rates differ significantly.
2. Calculate your true cost per message including infrastructure, an allocated slice of support hours, and a 5 to 8 percent failure buffer.
3. Choose the unit your customer thinks in. Some businesses think in “conversations,” some think in “contacts messaged,” some think in raw message count. Match your pricing language to how they already talk about the problem, not how the API technically works.
4. Model your price at three volumes: a small customer (say 1,000 messages a month), a mid customer (50,000), and a large one (500,000). If your margin collapses or turns negative at the top end, your pricing is broken before you’ve sold a single contract.
5. Set a monthly floor so tiny accounts don’t cost more to service than they pay you. A £0 platform fee with pure usage pricing sounds friendly but it means your support team is subsidising customers who barely use the product.
6. Test the price with five real prospects before you publish it. Not a survey, an actual conversation where you say the number out loud and watch their face. This tells you more in twenty minutes than a month of spreadsheet modelling.

Value-based pricing still applies here, even though it’s a commodity underneath

It’s tempting to price a messaging API purely on cost-plus math because the underlying product is a commodity. But the thing your customer is buying isn’t the message, it’s the outcome: fewer no-shows, faster payment collection, higher reply rates on a marketing campaign. This is exactly the value-based pricing principle that Philip Kotler spent decades writing about, and it holds up here better than most people expect. A salon that cuts no-shows from 18 percent to 6 percent using your reminder API isn’t comparing your £39 fee against a raw SMS rate, they’re comparing it against the revenue they were losing to empty chairs. Price against that outcome where you can, not just against your cost stack.

The pricing mistakes I see most often

• Copying a competitor’s public price without knowing their cost base. A bigger player can absorb thinner margins because of volume discounts from carriers that you don’t have access to yet.
• Ignoring country-by-country cost differences until a customer scales internationally and your margin quietly disappears. If your customer is expanding into new markets, it’s worth reading up on how researching a global expansion strategy works, because messaging costs are one of the most overlooked line items in that kind of planning.
• Underpricing out of fear of losing the deal. This is the one that Dan Kennedy hammered on for years in his direct marketing work, and it’s still true: undercharging doesn’t win you loyal customers, it wins you fragile ones who leave the moment someone cheaper shows up. There’s a good breakdown of that thinking in the business lessons from Dan Kennedy piece I wrote.
• No clear overage communication. If a customer gets a surprise bill because they didn’t understand what counted as a “message” versus a “conversation,” you’ll lose them even if your pricing was technically fair.
• Building the pricing page before testing the number. A polished pricing page with the wrong number just gets you fifty sign-ups you can’t service profitably.

Where AI-powered messaging changes the maths

If your messaging service includes any kind of AI-generated replies, smart routing, or automated conversation handling, your cost stack gets a fourth layer: the cost of the model call itself, on top of the carrier fee. This is where a lot of newer vendors get caught out, because they price purely on the messaging side and treat the AI layer as “included,” not realising a single AI-assisted conversation can cost several times more than the raw message delivery. Neil Patel has written extensively about how AI tooling changes cost structures across marketing tech, and the same logic applies directly here, worth a look in the business lessons from Neil Patel post if you’re building anything with an AI layer on top of your messaging API.

Practically, this means separating your pricing line items: base platform fee, per-message or per-conversation rate, and a distinct AI usage rate if applicable. Bundling them all into one flat number feels simpler for the customer but it hides exactly the cost variable most likely to spike unpredictably.

The number that should drive your decision

Forget market benchmarking for a moment. The single number that should drive your pricing decision is your gross margin target at your median customer’s volume, not your smallest or your largest. Aim for 60 to 75 percent gross margin on the messaging line if you want room to cover support, sales, and product development without needing every customer to be a whale. If your margin at the median customer is below 40 percent, you don’t have a pricing problem yet, you have a cost problem, and no pricing tweak fixes that until the underlying cost stack is sorted.

Frequently asked questions

Should I price per message or per conversation for WhatsApp specifically?

Per conversation almost always, because that’s how Meta bills you underneath, and pricing your customer differently from how you’re billed creates a mismatch that gets worse as volume grows; align your pricing unit with your actual cost unit wherever the underlying platform allows it.

What margin should I be aiming for on an API based messaging service?

Target 60 to 75 percent gross margin at your median customer’s volume, including carrier fees, infrastructure, allocated support time, and a delivery failure buffer of 5 to 8 percent, not just the raw carrier rate.

Is a free tier a good idea for a messaging API product?

Only if the free tier volume is small enough that your real cost per message stays under a few pounds a month per free account, otherwise you’re subsidising trial users indefinitely rather than converting them, and messaging costs (unlike most SaaS trials) don’t shrink to zero just because the account is free.

How often should I revisit my messaging pricing once it’s live?

Review it every time a carrier changes its rates, at least twice a year regardless, and immediately if you notice your largest customers by volume are also your lowest margin accounts, because that pattern almost always means your pricing model, not just the number, needs rebuilding.

Related reading: Plan to scale your service-based online business? Carry out these 10 tips to start with and How to Price Your Services When You Keep Undercharging.