Playbooks · 5 min read
Map the Journey From Real Messages, Not From a Whiteboard
Most customer journey maps are drawn in a room by people who are not customers. Somebody writes Awareness, Consideration, Purchase, Retention across a whiteboard, arrows get added, and the artefact goes into a deck where it is never contradicted by anything.
There is a version of this exercise that is genuinely useful, and it is barely the same activity. It is closer to reading a stack of transcripts with a pen. On a channel where every interaction is a private thread, the material for it is sitting in the inbox already.
Why this channel makes the exercise unusually honest
A journey map for a website is stitched together from analytics — sessions, sources, page paths — and each stitch is a guess about intent. A journey on a messaging channel is a conversation with words in it. You do not have to infer why somebody came. They typed it.
That is a rare advantage and it is wasted almost everywhere, because the inbox is treated as a task queue to be emptied rather than as a record to be read.
The exercise
Set aside two hours. Take the most recent two hundred conversations — not a selection you remember, the actual scroll, including the ones that went nowhere.
- Who spoke first, and how did they get here. Did they message you, or did you message them? If they messaged, is there a prefilled string or a first line that tells you which door they came through — a link, a QR, an ad, a number from an invoice, a friend's forward?
- What the first ask was, verbatim. Copy the actual words. Not "pricing enquiry" —
ye kitne ka hai bhaiya. The verbatim column is the one you will use twice, and the summary column is the one you will throw away. - What ended the thread. They bought. They said thanks and stopped. They asked something you never answered. Somebody on your side stopped replying. It just trailed off. Be honest about the last two, because they are the largest categories in most files and nobody records them.
- How long the first reply took. Minutes, hours, or the next morning. One column, a rough number.
Then sort by column two, and read.
What you will find, roughly in this order
The map is not a funnel. Almost nobody moves through neat stages. What you have is a small number of entry points, each producing a characteristic first question, and a small number of endings. Drawing it as a line with arrows is already a distortion.
The entry point predicts almost everything. Somebody who arrived from a product page asks about that product. Somebody who scanned a QR at the counter is standing in front of you. Somebody who got the number forwarded by a cousin has no idea what you sell. These are three different conversations and treating them as one funnel stage is why generic first replies underperform.
One question is more common than you thought and one is more expensive than you thought. The common one is the automation candidate. The expensive one — the thread that runs fourteen messages every time — is usually a missing page on the website or a policy nobody has written down.
The dead ends have a pattern. Threads that trailed off usually did so at the same point, and it is almost always a message from your side that ended with a link or a statement rather than a question.
The most useful column is the last one. Not how conversations start — how they stop.
Turning the map into something built
The temptation is to draw the finished map as a flow with branches and then look for a tool that renders it. Resist it. What the map actually produces is a short list of independent things:
What the map hands you
What it does not justify
Each of those left-hand items is a small, separate automation that knows nothing about the others. Together they read to a customer like a business that has its act together, which is what a journey map was supposed to produce all along.
Re-run it, and keep the verbatim column
Do this quarterly, or after anything changes — a new product, a new ad, a new payment method. The verbatim column from the last round is the input for the next one: new spellings show up, old questions disappear because you fixed the page that caused them, and a question you have never seen before starts appearing three times a week.
That drift is the real finding. A journey map is not a document. It is a habit of reading what customers actually typed, and the document is a by-product.
What not to record
A word on restraint. You are reading real conversations that people had with a business in private, and the map does not need names, numbers, or the contents of anything personal. Code the pattern, not the person. What data protection obligations apply to how you store and analyse those threads depends on facts about your business, and it is worth taking advice on rather than assuming.
Where this stands today
PostEngage does not ship WhatsApp. On Instagram it answers comments and DMs, captures leads from those conversations, and exports them as a CSV. There is no analytics product here and no journey dashboard — the exercise above is done by a person with a spreadsheet, which is also the only way it has ever worked.
For the Instagram version of the same read-the-threads discipline, the customer journey post is here. For what to do with the four questions the map hands you, support automation is the next step, and keyword lists get harvested the same way.
