Consent, Frequency Caps and Message Fatigue for Indian D2C
- The common failure is a single boolean on the customer record labelled subscribed.
- Per campaign caps fail predictably. Every campaign owner sets a reasonable limit for their own flow.
- Unsubscribing takes effort. Most fatigued customers do something cheaper first.
Most Indian D2C brands discover message fatigue the way you discover a slow leak. Not through an alarm, but through a slow decline in performance that gets blamed on creative, on the algorithm, on the season. The list is not smaller. It is just quieter. And by the time the unsubscribe rate moves, the damage is a quarter old.
Fatigue is a measurement problem before it is a strategy problem. Here is how to instrument it.
Hold consent per channel and per purpose
The common failure is a single boolean on the customer record labelled subscribed. It cannot answer the questions that matter. This person accepted order updates on SMS, did they accept promotional email. They opted in through a giveaway two years ago, does that cover a subscription reminder today.
Model consent as a grid. Channel across one axis: SMS, email, push, and any messaging surface you use. Purpose across the other: transactional and service, lifecycle and account, promotional and marketing, research and surveys. Each cell holds four fields. Status, timestamp, source, and the exact wording shown at capture.
Source and wording are what make this defensible. Twelve months later, when someone asks why a contact received a campaign, you can produce the checkbox they ticked and the words next to it. A boolean cannot do that.
Two operating rules follow. Consent for one purpose never implies consent for another, so a transactional relationship does not authorise promotional sends. And consent for one channel never implies another, so an email opt in does not license SMS. Indian consent and communication rules span multiple frameworks including telecom preference rules and evolving data protection obligations, and they continue to change. Confirm current requirements with counsel rather than working from a general summary.
Global caps beat per campaign caps
Per campaign caps fail predictably. Every campaign owner sets a reasonable limit for their own flow. Nobody sets a limit on the sum. The customer who is simultaneously in a winback, a replenishment reminder, a category launch and a cart sequence receives eleven messages in a week, and no individual owner did anything wrong.
The cap has to live above the campaigns, enforced at the send layer, and it has to be a hard block rather than a warning. Structure it in three tiers.
A global cap on total marketing contacts per person per week, counted across every channel together. A per channel cap underneath it, because inbox tolerance and SMS tolerance are different. And a priority ladder that decides who wins when the cap is hit, so the lower value send is dropped or deferred rather than the high value one.
Exempt genuine transactional messages from the cap. Order and refund updates are owed regardless. Be strict about that boundary, because the fastest way to break a cap system is to relabel marketing as transactional to escape it.
Quiet hours are the other half. No marketing sends late at night or early morning, held in the recipient’s local context. Batch anything generated during quiet hours to the next allowed window rather than dropping it silently.
Unsubscribe rate lags, engagement leads
Unsubscribing takes effort. Most fatigued customers do something cheaper first. They stop opening. They stop tapping. They mark as read without looking, or they let the mailbox provider start filing you away. By the time unsubscribe rate rises, you have been over sending for weeks and your inbox placement has already softened.
Watch the leading indicators instead, all tracked as trends per segment rather than as single period snapshots.
Open or read rate on a rolling thirty day basis, by tenure cohort, so a healthy new cohort does not mask a decaying old one. Click to open rate, which separates a subject line problem from a content problem. Time to first open, which lengthens before it stops. The share of your active list that has engaged at all in ninety days. Spam complaint rate, which is small in absolute terms and violent in consequence. And send to engagement ratio, meaning how many messages it now takes to produce one engaged action compared with a quarter ago.
That last one is the honest fatigue metric. When it doubles, you are getting the same result by shouting louder, and the cost of that shouting is being paid in deliverability.
List hygiene and sunsetting
Continuing to mail people who never open is not neutral. Mailbox providers weight engagement heavily, so a large unengaged tail drags placement for the engaged core. You are trading the customers who want to hear from you for the ones who do not.
Run a sunset ladder rather than a sudden purge. Reduce frequency sharply for contacts inactive past ninety days. At one hundred and eighty days run a short reengagement sequence, two messages, with an explicit choice to stay. Non responders move to suppressed but retained, meaning excluded from sends while still counted for analytics and still eligible for transactional messages if they order. Reactivate only on a genuine action, not on a passive open.
The commercial case is arithmetic, not sentiment. A list of two lakh with sixty percent engaged produces more revenue and better placement than a list of six lakh with twelve percent engaged, and it costs less to send to. Report engaged list size as the headline number and total list size as a footnote. Teams optimise the number you put on the slide.
A monthly audit routine
Ninety minutes, same date each month, one owner.
Pull last month’s sends per person and plot the distribution. Look at the ninety fifth percentile, not the average. If a slice of customers is receiving three times the median, your caps are not being enforced where it counts.
Take twenty random contacts and reconstruct every message they received, in order, on every channel. Read it as they did. This finds the collisions no dashboard shows, such as a winback landing the day after a purchase.
Sample fifty recent opt ins and verify each one has a source and captured wording. Any that do not are a gap in both compliance and diagnosis.
Review the leading indicators as trends over six months, not against last month. Check quiet hour breaches and cap overrides, and require a written reason for each override. Then confirm the sunset ladder actually ran, because suppression jobs fail quietly and nobody notices for two quarters.
Close with one number reported to the business: engaged contacts, defined as anyone who took an action in the last ninety days, tracked month over month. If that number is growing while your total list is flat, the programme is healthy. If total list is growing while engaged contacts are flat, you are buying decay.