EVERY TOUCH, IN ORDER

Ask five people how many messages a new customer gets in their first two weeks and you’ll get five answers. Messages come from the store platform, the email and SMS tool, the tracking app, the review app and the helpdesk, each set up by someone thinking about their own. Only the customer sees the whole sequence.

Two lists

List what you send: order and shipping notices, flows, campaigns, SMS, tracking-app notices, review requests, subscription reminders. Then what customers start: tickets, chats, replies, return requests, reviews, cancellations. In Lemon and Verhoef’s terms, the first list is mostly brand-owned and partner-owned; the second is where customer-owned touch points leave a trace Published.

For each touch, record its stage, the day it fires (from the first order), the trigger and the system that sends it.

Lay it on one timeline

Then pull the actual message log for a handful of real first-time customers and lay it out by day. Picture a customer who orders on a Friday, during a sale week, and whose parcel is delayed:

Made-up example. Day 0 is the first order. Day 3: a delay notice, a flow email, a sale campaign and a sale SMS. Delivery on day 6; review request on day 7, before most customers have used the product. Days 8 to 14: nothing.

Two patterns stand out.

Collision days

A collision is a day when messages from different systems pile up, or when a marketing message lands next to a problem. The delay notice on day 3 is the worst kind: the customer learns their order is late and, an hour later, gets a text urging them to buy more. Each system behaved correctly. Nobody owned the day.

Three rules prevent most collisions:

• A problem suppresses marketing. A delay, a failed payment, an open ticket or a return in progress pauses campaigns and sales SMS for that customer for 48 hours, or until it’s resolved.
• A daily cap across systems. No more than two marketing messages a day per customer, counting email and SMS together.
• One calendar owner. One person approves every new automated message, from any tool, against the timeline.

The Whole Machine covers flows, frequency and deliverability. The point here is narrower: a collision is a journey problem, and you only see it when every system is on one timeline.

Silent stretches

The opposite problem sits at days 8 to 14: after the review request, nothing, until a replenishment email weeks later. That’s the stretch between delivery and first use, the stage chapter 5 found running furthest past its target. Silence isn’t always wrong. But it’s usually where the next question goes unanswered.

The Corporate Executive Board’s contact-center research, published in HBR in 2010, is useful here. Matthew Dixon, Karen Freeman and Nicholas Toman reported that 22% of repeat calls involved “downstream issues related to the problem that prompted the original call”, and that “fifty-seven percent of inbound calls came from customers who went to the website first” Reported. That’s the firm’s own research, from service centers, but the lesson carries: the next question is predictable. In a DTC journey it’s how to start, what to expect in week one, and when to reorder. Answer it before it’s asked.

What customers start is a stage signal

Tag every ticket with the stage the customer was in when they wrote, which most helpdesks can do from order status, and report contacts per 100 orders by reason and stage. A rise in how-do-I questions from customers without a first-use signal tells you more about that stage than any survey. Where-is-my-order contacts belong to The Kept Promise; returns to The Return Trip.

Do this

• Export the full message log for ten customers who first ordered last month, from every tool, and lay it out by day as above. Circle every day with more than two marketing messages or any marketing next to a problem notice. Mark every stretch of seven days or more with nothing useful in it.