--- title: "Conversational marketing, minus the hype" summary: "A useful definition has to exclude something. This one — the customer's reply changes what happens next — excludes most of what gets sold under the name, and tells you how to staff and measure the rest." canonical: https://cpaas.co/en/learn/conversational-marketing-defined kind: explainer depth: foundation origin: original published: 2026-09-09T16:26:18.899Z updated: 2026-10-14T16:26:19.457Z topics: ["Conversational Marketing", "Fundamentals", "Customer Experience"] language: en --- # Conversational marketing, minus the hype The term is used loosely enough to mean "marketing, but on WhatsApp". That definition excludes nothing, so it is useless. ## A definition that excludes something Conversational marketing is marketing where **the customer's reply changes what happens next**. Example: a furniture shop messages a customer a week after delivery: "How is the new sofa? Reply 1 for care tips, 2 if something is wrong, 3 to see matching cushions." A "2" opens a support case with a person; a "3" sends three options with photos; silence ends the sequence. That is conversational. If the same three messages go out on the same schedule whether or not anyone replies, that is **broadcast marketing delivered over a messaging channel**. Nothing wrong with it — but calling it conversational will lead you to staff and measure it wrong. ```diagram:decision # A week after the sofa arrived Did the customer reply "1"? | Send the care tips | Did they reply "2"? | Open a support case with a person | Did they reply "3"? | Send three matching cushions, with photos | Silence ends the sequence > The reply decides what happens next. If all three messages would have gone out anyway, it is a broadcast. ``` ## What follows from the definition **A reply path is mandatory.** Sending from a number or sender name that cannot receive replies rules the programme out immediately. Obvious, and violated constantly. **Somebody has to be there.** A person or an automated agent must handle what comes back, within a time the customer finds reasonable. A conversation you do not answer is worse than a broadcast you never sent, because you asked. **Branching is the product.** If the journey has no branches, the "conversation" is decoration. ```diagram:matrix Broadcast on a messaging channel | Same messages, same schedule, whether or not anyone replied. Nothing wrong with it — just measure it as a campaign. Conversational | A reply changes what happens next. Branching is the product, not a feature of it. Answered by nobody | A conversation you invited and cannot staff. Worse than a broadcast, because you asked. Answered badly | An agent that traps the customer in a menu. Worse than no reply, because it used up their attempt. > Only the second quadrant is conversational marketing. The other three are what the term often describes in practice. ``` ## Why it works, when it works Three plain mechanisms: - **People read these channels.** Personal messaging apps hold attention that email has largely lost. - **Replying is easier than clicking through.** "Yes, tell me more" costs one word instead of a page load and a form. - **A reply is a strong signal.** On WhatsApp it opens a 24-hour window in which you may reply freely, in your own words (see [the 24-hour window](/en/learn/the-24-hour-window)). In most regimes it is also good evidence of interest — but it is not, on its own, consent to future marketing. ## Operations before tools The usual failure is not the software. It is starting conversations the organisation cannot finish. Before launch, answer: 1. **Who answers, and in which hours?** 2. **What response time do we promise, and is it written where the customer can see it?** 3. **What happens outside those hours?** An honest "we'll reply tomorrow morning" is fine; silence is not. 4. **When does an automated agent hand over to a person, and does the person see the history?** A programme with no answers here is a broadcast that has accidentally given people a way to reply. ```diagram:checklist # Answer these before launch [ ] Who answers, and in which hours? [ ] What response time do we promise? | And is it written where the customer can see it? [ ] What happens outside those hours? | An honest "we'll reply tomorrow morning" is fine; silence is not [ ] When does an agent hand over to a person? | And does the person see the history? > Four open boxes and a reply path: a broadcast that has accidentally let people answer. ``` ## Frequency is the whole discipline The channel's advantage is that it feels personal. That advantage is spent, not renewed. On email, over-sending erodes engagement slowly; on messaging apps, people block — and a block is permanent and invisible to you. Blocks and spam reports can also hurt your sender's quality rating and limits on some platforms. Practical guardrails: - a **frequency cap** across all journeys, not per journey; - a **minimum gap** between promotional messages to the same person; - a rule that a **service message never carries a promotion**. Mixing them is how a trusted channel becomes a muted one. ## Measure the conversation, not the campaign Opens and clicks miss the point. Measure: | Metric | Why | |---|---| | Reply rate | The only number that tells you it felt like a message | | Resolution rate | For anything that started as a question | | Time to first human response | What people remember | | Opt-out and block rate per send, as a trend | The early warning of over-use | | Outcome per thousand sent, against a holdout | Whether it changed anything | See [measuring a conversational programme](/en/learn/measuring-a-conversational-programme) for the detail. ## Tell people when they are talking to AI If an automated agent answers, say so. Since **2 August 2026**, Article 50 of the EU AI Act requires that people in the EU be told clearly, at the first interaction, when they are interacting with an AI system — and the disclosure cannot be buried in terms and conditions. Other places have their own rules; California, for example, has required disclosure of bots used to sell to people in some contexts since 2019. Disclosing everywhere is simpler and builds trust. See [disclosing AI in a conversation](/en/learn/disclosing-ai-in-a-conversation). ```diagram:timeline # Disclosure rules, by date 2019 | California | Disclosure of bots used to sell to people, in some contexts 2 August 2026 | EU AI Act, Article 50 | People in the EU told clearly, at the first interaction, that they are dealing with an AI system ``` ## Where to go next - [Designing a journey that branches](/en/learn/designing-a-journey-that-branches) — the next step in this track. - [Measuring a conversational programme](/en/learn/measuring-a-conversational-programme) — the deeper level. - [WhatsApp template strategy](/en/learn/whatsapp-template-strategy) — the first message, done well. ## Sources [^1]: [Commission starts enforcing AI Act rules and new transparency requirements on 2 August](https://digital-strategy.ec.europa.eu/en/news/commission-starts-enforcing-ai-act-rules-and-new-transparency-requirements-2-august) — European Commission [^2]: [Regulation (EU) 2024/1689 (Artificial Intelligence Act)](https://eur-lex.europa.eu/eli/reg/2024/1689/oj) — EUR-Lex [^3]: [SB-1001 Bots: disclosure](https://leginfo.legislature.ca.gov/faces/billNavClient.xhtml?bill_id=201720180SB1001) — California Legislative Information, 2018 --- Published by the company behind five products in this category. Those products appear in this site's directory alongside competitors under the same published criteria and are labelled as its own. Editorial content does not recommend them.