Twilio Alternative: Which Job Did You Hire It For?

Two categories of product both say "WhatsApp" on the tin, and they are bought for different reasons. Getting the category right before the feature comparison saves a quarter.

A communications API — the category Twilio, Vonage and MessageBird sit in — is the right purchase when the sending has to happen inside software you are already building, on your own schedule, under your own logic. What you buy is programmable access across channels. What you then own is everything around it: the list, the consent record, the templates, the scheduling, the inbox and the reporting. If you have engineers and a product roadmap that includes messaging, that is a reasonable trade.

A Shopify-native messaging tool is the right purchase when the problem is that WhatsApp is not earning anything yet and nobody on the team writes code on Tuesdays. What you buy is event-driven sending — a message that goes out because a checkout was abandoned or an order shipped — plus the list to send it to and the attribution to prove it worked. That is what Keaz is: ten flows and nine growth tools, pre-written and pre-configured.

Here is the honest boundary. Keaz does not do voice, it does not route arbitrary SMS, and it is not where you send one-time passcodes. If any of those is the job you hired an API for, no WhatsApp marketing tool replaces it — and the right answer may be to keep both, with the API doing the infrastructure work and the marketing channel doing the marketing.

A send endpoint is one piece of a channel. These are the other five, and in an API-first setup every one of them is yours to build and maintain.

1. Collecting phone numbers and recording consent. Not a field in a database — a checkout configuration, a post-purchase prompt, a widget, and a stored record of what each person agreed to and when.
2. Template lifecycle. Drafting, submitting for approval, handling rejections and re-categorisations, and keeping the approved versions in sync with what your code sends.
3. Segmentation and scheduling. Deciding who gets which message when, and not sending it twice because two services both fired.
4. An inbox. Customers reply. Somebody has to see the reply, inside the window where a free-form answer is still allowed, with the clock visible.
5. Attribution. Not "we sent 4,000 messages" but "Cart Recovery earned this much last month" — which means a per-flow mechanism, not a channel-level total.

None of this is an argument against APIs. It is an argument for counting the build before comparing anything, because the five items above are where the quarters go. If you are at the earlier question of who can legitimately put you on the platform at all, how to choose a WhatsApp Business Solution Provider covers the shape of that decision. On our side the access path is a guided one — connecting a WhatsApp Business account to Keaz runs through 360dialog and the Meta Business Manager in about fifteen minutes of active work, plus Meta's own review time.

This is the part of the comparison that no vendor can change, and it is worth knowing before you shortlist, because it bounds what any of them can do for you.

• Opt-in comes first. Meta requires an opt-in before you message anyone, and it has to name the business. The collection method is yours to choose; the obligation is not.
• The 24-hour customer service window decides message type. A user message opens it and each further user message resets it. Inside it you can answer freely; outside it only approved templates go out.
• Volume is capped and earned. A newly created business portfolio starts at 250 unique recipients per moving 24-hour period outside that window, rises to 2,000 by completing a scaling path, and then to 10,000, 100,000 and Unlimited through automatic scaling when message quality is high and usage thresholds are met.
• The limit is portfolio-level, not per number. It is shared across every business phone number in the portfolio, so a second number does not buy a second allowance.

Those four rules come from Meta's own documentation (Meta, Messaging Limits, retrieved 9 November 2026). Rates change on Meta's own schedule and are not reproduced here; the dated figures live on our WhatsApp platform reference, each with the date it was retrieved. Any vendor quoting you a per-message price is quoting Meta's price plus their own margin — check the underlying rate yourself.

Feature lists rot and every vendor writes their own. These five are answerable from a trial account in an afternoon, and they still matter a year later.

1. Does it read your shop, or only send into it? The test: can a message be triggered by an abandoned checkout or a completed purchase without exporting anything? In Keaz those are flow triggers alongside chat-in and webhook triggers, and connecting Shopify is a click, not a project.
2. Who operates it in week six? Setup is a one-off; the weekly routine is not. If the answer is "an engineer", that engineer is part of the price, and their time is the most expensive line in the comparison.
3. Does it grow the list, or only message it? A tool that sends beautifully to 40% of your customers is worth less than a plain one that reaches 85%. Check for a phone field at checkout, a post-checkout capture widget and a documented way to re-permission dormant customers.
4. Can you see revenue per flow, not just per channel? Keaz creates one discount code per flow in Shopify automatically and attributes redeemed revenue to that flow. Ask any candidate how it answers "what did Cart Recovery earn last month".
5. Can you model the outcome before you commit? Any vendor can show a case study. Ask for a number computed on your shop: the Forecast runs on your own Shopify data, and the per-industry ranges show what the first 90 days look like in your vertical.

One more, if you came to an API for SMS rather than WhatsApp: the channel question is separate from the vendor question. RCS, SMS and WhatsApp compared for a shop settles that one first, and it changes the shortlist.

With a number you did not have to argue about. Two scores and one attribution mechanism produce it.

Customer Contact Coverage (CCC) is the share of your last-12-months Shopify customers you can reach on WhatsApp at all. Stores typically start between 35% and 55%. Getting past 85% takes about 60 days and is mostly plumbing, not persuasion — and no amount of integration quality compensates for a low number here.

Marketing Contact Coverage (MCC) is the share of your email subscribers also reachable on WhatsApp. It usually starts at 15–30% and settles at 50–70%, and it tells you whether the channel is adding audience or re-messaging the same people. On our own audience data, 24% of customers are reachable by email only, 38% by both, 34% on WhatsApp only and 3% on neither — which is why WhatsApp complements email here rather than replacing it.

Published benchmark bands put WhatsApp read rates above 90% against email open rates of 30–40%, with roughly 1.5× the conversion rate on a shared audience. Those are bands, not promises — how we calculate them is written up separately. Consent underpins all of it: the GDPR frames it as freely given, specific, informed and unambiguous (Regulation (EU) 2016/679, 2016), and a migration does not carry it across.

• Comparing endpoints before naming the job. Two products in different categories both list "WhatsApp" and "automation". The comparison means nothing until the problem you are paying to remove is written down.
• Pricing the messages and not the build. The per-message line is the visible cost. Opt-in capture, template lifecycle, an inbox and attribution are the invisible one, and they are larger.
• Replacing an API that is doing a job nothing else does. Voice, one-time passcode delivery and arbitrary SMS routing do not move to a WhatsApp marketing tool. Split the account rather than forcing one vendor to cover both.
• Migrating a contact list and treating consent as migrated too. The data moves; the permission does not follow it. Re-check the opt-in level before the first send.
• Expecting a new vendor to lift your messaging limit. The limit sits on the business portfolio and is earned through quality and volume, not bought.
• Evaluating on a test send instead of on coverage. A test message always looks good. Run the coverage number first; it sets the ceiling on everything after it.