Campaigns

Everything else in Tola is about answering customers well. Campaigns are the one place you start the conversation — an order update, a renewal reminder, a nudge to someone stuck at checkout. Used properly this is the highest-return thing in the product. Used carelessly it is how you get your WhatsApp number blocked.

Two kinds of campaign

One-off campaigns

Create one under Campaigns, pick the channel, choose your audience, write the message and schedule it.

Good uses, roughly in order of how well they perform:

• Order and delivery updates. Highest engagement of anything you will send, because it is information the customer wants.
• Appointment and booking reminders. Measurably reduces no-shows.
• Renewal and payment reminders. Recovers revenue that would otherwise churn silently.
• Service notices — an outage, a change of hours, a policy change.
• Genuinely relevant offers, to a narrow segment, occasionally.

That last one is where campaigns go wrong. A broadcast to your entire contact list is not a campaign, it is a mailshot, and the channels it goes out on punish it.

WhatsApp campaigns

WhatsApp is the highest-value campaign channel and the most tightly governed. Two rules decide everything:

You need an approved template. You cannot send a freeform message to someone outside the 24-hour window, which by definition is everyone in a campaign audience. Templates are registered with Meta in advance and reviewed before approval. See WhatsApp.

Templates are personalised from your contact data. One approved template — "Hi name, your order number is out for delivery" — becomes a personal message for every recipient, filled automatically from your contact records. This is why keeping custom attributes accurate pays off.

Practical guidance:

• Send utility messages, not marketing, wherever the message can honestly be framed as one. Order updates, reminders and notices are welcomed; promotions are tolerated at best.
• Segment tightly. A hundred people who want this message beat ten thousand who do not.
• Honour opt-outs immediately and visibly. Someone who asks to stop and receives another message will report you.
• Start small. Send to a few hundred, watch the reaction, then widen.

Ongoing live chat campaigns

These trigger on your website: a visitor is on a page for a set time and receives a proactive message.

They work when they are specific and helpful. On a pricing page: "Not sure which plan fits? Tell me your team size and I'll say." At an abandoned checkout: "Anything blocking you? I can help." On a documentation page people get stuck on: an offer of the answer.

They fail when they are generic and everywhere. "Hi! How can I help you today?" on every page of the site is the thing that trained a generation of people to close chat widgets on sight.

Pick two pages where a stuck visitor is genuinely likely. Leave the rest alone.

Choosing an audience

Audiences come from your contact data — labels, segments and custom attributes. See contacts and companies.

Use a segment rather than a fixed list wherever you can. A segment is a rule, so it stays correct: "customers whose renewal date is within thirty days" is right every time you use it, while a spreadsheet exported last month is already wrong.

Before you send, check three things: that everyone in the audience should genuinely receive this, that nobody in it has opted out, and that the personalised fields are actually populated. A message reading "Hi ," goes out to more customers than anyone would like to admit.

Replies come back to the inbox

A campaign is not a broadcast into nowhere. When someone replies, it becomes an ordinary conversation in your queue, assigned and answered like any other.

Plan for that. A campaign to five thousand people with a two percent reply rate is a hundred conversations arriving at once. Send when you have the staffing to answer, not at 5pm on a Friday.

This is also the strongest argument for sending fewer, better campaigns: every message you send is a potential conversation your team has to have.

Measuring a campaign

Look at three things:

Did it get delivered? Delivery failures point at data quality — wrong numbers, stale addresses.

Did it cause replies? For a service message, few replies is success. For a nudge, replies are the point. Know which you sent.

Did it cause the thing you wanted? Fewer missed appointments, more renewals, fewer abandoned checkouts. This is the only measure that matters, and it is the one people skip because it is the hardest to attribute.

Watch the negative signals as closely as the positive ones. Opt-out rate and block rate are the early warning that you are sending too much, and they show up long before the number is restricted.