Omnichannel CRM for Ecommerce

Here's a morning in the life of a customer at a brand with five well-run channels. At 8am she gets a win-back email, because the email tool has decided she's lapsed. At 8:40 a push notification arrives about a sale, because the push tool doesn't know about the email. At noon the loyalty platform texts her about expiring points. At 2pm the subscription platform emails a shipment reminder.

Every one of those messages is defensible. Together they're a brand that seems to have no idea who she is — and she has already unsubscribed from one of them.

That's the omnichannel CRM problem. It isn't a shortage of tools. It's five tools with no shared idea of the customer.

The collision problem

Independent channel tools each make locally sensible decisions with locally available data. The email platform doesn't know a push went out an hour ago. The loyalty app doesn't know the customer is mid-cancellation in the subscription tool. Nothing is misconfigured; there's just no layer above them arbitrating who gets to speak.

The symptoms are recognizable: message pile-ups on big days, contradictory offers, win-back campaigns aimed at customers who bought yesterday, and opt-out rates that climb faster than list growth. The failure isn't in any one channel — it's in the absence of coordination between them.

The single customer view comes first

Every fix starts with one profile per person, holding purchases, subscription status, loyalty tier and points, channel consent, engagement recency per channel, and browse behavior — all resolved to one identity across email address, phone number and device token.

Without that, "omnichannel" is just a word for owning several tools. With it, you can ask the only question that matters before any send: given everything we know about this person right now, is this the right message on the right channel today?

Pick a control layer

Someone has to be in charge. In practice, the platform holding the richest customer data and the most sophisticated segmentation — usually the ESP — becomes the control layer, and other channels become executors that fire when the control layer says so.

That's a deliberate architectural choice, not an accident, and it's what stops five roadmaps from each optimizing their own numbers. It also means integrations get judged on one criterion: does this tool write its events back to the control layer, and can it be triggered by it? A tool that can't do both will always drift out of coordination.

Suppression rules that cross channels

Once there's a control layer, cross-channel suppression becomes possible — and it's where most of the customer-experience gain lives.

The rules are unglamorous and effective. Anyone in an active subscription save flow is suppressed from promotional sends, because a discount arriving mid-cancellation reads as an insult. Anyone who bought in the last 48 hours is suppressed from cart and browse abandonment. A customer who received a launch SMS this morning doesn't get the launch email this afternoon. Loyalty point reminders don't ship on the same day as a promotional push.

Each rule removes a specific collision you can name — which is the test for whether it belongs.

Frequency caps that count everything

Most brands cap frequency per channel: four emails a week, two texts. A customer on all five channels can still receive fifteen messages while every individual cap is respected.

A shared cap counts total brand contacts per person per week, across all channels, with priority rules for what gets dropped when the ceiling is reached. Transactional and service messages sit outside the cap; promotional ones compete for the remaining slots. It's the single change that most reliably lowers opt-out rates without lowering revenue, because what gets cut is the marginal message that was never going to convert.

Channel of choice, and channel of moment

Two ideas do the sequencing work.

Channel of choice is where a specific customer actually engages — some people never open email but respond to every text; some ignore push and read every newsletter. When the profile shows a clear preference, lead with it.

Channel of moment is what the message needs. Time-sensitive and short goes to SMS or push. Depth, storytelling and merchandising go to email. Status, tiers and rewards live in loyalty, surfaced by the others. Shipment and billing events belong to subscriptions, escalated by SMS when something breaks.

When the two conflict, the moment usually wins for urgency and the preference wins for everything else — the same segmentation discipline we apply within a single channel in ecommerce email segmentation, extended across all of them.

What it looks like when it works

Our Centr case study is this problem at full scale: a fitness membership brand running email, SMS, push, loyalty and subscriptions, where CRM had to work as one system rather than five. Rebuilding lifecycle segmentation across acquisition, win-back, paying members and free trials, then aligning email, push and in-app messaging under it, produced 2.5x higher engagement year over year during key sale periods, more than $1.9M in CRM-attributed revenue across the BFCM season, and a 20% lift in trial-to-paid conversions from improved onboarding.

The onboarding number is the tell. Trial-to-paid moved because the messages a new member received stopped competing with each other.