One cap for everything
Many CRM setups manage contact pressure with a single rule: no more than a set number of messages per channel and week. That is better than having no rule at all. But it is too blunt.
This cap treats four very different customers the same way. One has explicitly clicked “not interested”. One complained yesterday. One has just bought something. And one has not opened anything for months.
Three scenes, all within the cap.
An energy supplier’s customer disputes their annual bill on Monday. On Wednesday they receive the newsletter promoting the new green electricity tariff.
A customer buys an espresso machine from an online shop. Two days later the shop advertises espresso machines to her.
A mobile customer reports a fault on their line. The next email recommends a tariff upgrade.
None of these messages broke a rule. Each of them costs trust.
LinkedIn has measured how expensive too much contact becomes. In an experiment by the LinkedIn engineering team, one group of members received every planned email, while in a second group emails were randomly dropped. The full group showed 45% more negative reactions to the emails. The reduced group, in return, viewed only 2.6% fewer pages.
So dropping emails cost a little usage. The full volume caused a lot more rejection.
Rejection affects deliverability. According to Google, senders who send more than 5,000 messages a day to personal Gmail accounts should offer one-click unsubscribe in marketing emails. The spam rate should stay below 0.1% and never reach 0.3%. If you send too much, or send things that do not fit, your emails risk landing in the spam folder. That includes the ones the customer wants.
Five principles for contact rules that tell signals apart
Good contact management asks a second question before every message. Alongside “how often?” comes “what happened last?”. The following five principles help you answer it.
1. The strength of the signal sets the length of the pause
An explicit no is the strongest signal a customer can send. It deserves the longest pause, and only for the topic that was declined. Someone who states “not interested in TV packages” in their mobile provider’s customer portal should hear nothing about it for much longer than after any other signal. The channel stays open for their mobile contract.
A complaint is also a strong signal, but a different one. Here a pause that is not tied to a date fits better. It ends when the case is resolved. The energy supplier from the first scene would need a single field for this: complaint open, yes or no.
A purchase is a positive signal and still a reason to pause. The pause applies to the same product category and lasts as long as its purchase cycle. Someone who buys coffee beans will soon need more. Someone who buys an espresso machine will not look for a new one for a long time, though accessories fit right away.
Silence is the weakest signal. Someone who does not open has not declined anything. They receive less mail but stay reachable.
2. Every brake needs an amplifier
Most rule sets only know the brake. Yet customers also send positive signals. Someone clicks on a guide, looks at a product several times or asks support a question.
An amplifier is the counterpart to the pause. If a SaaS customer clicks on the article about the new interface, the topic moves up for a while. After that the signal slowly loses weight, just as the pause after a no eventually runs out.
A rule set without amplifiers knows when to stay quiet. It does not know when to speak.
3. Expensive channels pause longer
Channels cost different amounts, to you and to the customer. An in-app notice costs almost nothing and is quickly forgotten. A phone call ties up staff and interrupts the customer in the middle of their day. Push notifications, email and letters sit in between.
The rule: after a negative signal, the most expensive channel pauses longest. Phone calls and letters wait longer than the digital channels. An in-app notice may return earliest. If a SaaS customer has declined an upgrade email, sales should not call them shortly afterwards about the same topic. A discreet notice in the product may come back sooner.
4. One suppression list for all systems
Without a shared rule, exceptions pile up campaign by campaign. The win-back flow excludes onboarding, the newsletter excludes the current promotion. With every new flow this list grows longer, and at some point nobody can keep track of it.
A central suppression list is simpler. Every system that selects recipients reads it before sending. It records who is currently paused for which topic on which channel, and until when. This applies to the newsletter tool just as much as to sales, the customer survey and the service provider that sends post on your behalf.
For addresses, this has long been standard in many platforms. Adobe Journey Optimizer, for example, collects undeliverable addresses and spam complaints in a suppression list and automatically excludes them from journeys and campaigns. A suppression list by signal carries this principle over to topic and channel.
Don't forget the customer survey. To the customer, it is just another contact.
5. Service counts separately, hard blocks still apply
The notice of a price change from the energy supplier, the payment reminder in a SaaS product, the shipping confirmation from the shop: messages like these should not count against the marketing cap. Otherwise an invoice pushes out the next newsletter. Or, worse, a newsletter pushes out the invoice.
Many tools already separate the two. According to the Klaviyo Help Center, its Smart Sending feature by default lets no further marketing email through to the same person for 16 hours after a marketing email. This applies only to marketing messages. Transactional emails are not blocked and do not trigger the block.
A few blocks do apply to service as well. An undeliverable address stays undeliverable, even for the invoice. Good rule sets also make sure that anyone who has unsubscribed from marketing does not get it slipped into a service email.
The self-test for your next CRM meeting
You can find out whether your setup tells signals apart without starting a project. Sit down with the person responsible for your CRM and go through these questions. Ask to be shown each answer in the system.
- What happens when a customer explicitly says they are “not interested” in a topic? How long does that apply, and for which channels?
- What happens after a complaint? How does the system recognise that the case is resolved?
- Does a customer receive marketing for exactly what they have just bought?
- Which rule responds to interest, and how long does it last?
- How many systems select recipients? Do they all read the same suppression list, including surveys and sales?
- Which messages count as service, and who decided that?
- Does every rule have a responsible person and an expiry date?
For each question, note one of three results: covered, partly covered, open. A good answer points to a field and a rule and names a person. If an answer only points to the tool, ask which setting in that tool carries the rule.
Start with question 1. It shows most directly whether your rule set distinguishes between customers or only counts messages.
What standard tools provide and what is missing
Your marketing automation system usually already provides the basics. Check whether these functions are available and switched on:
- caps per channel and time window
- unsubscribes and handling of undeliverable addresses
- quiet hours, for example no push notifications at night
- rules on whether a customer can go through a flow more than once
- the separation of marketing and transactional messages
While you are at it, check one detail: how your tool counts a time window. Some systems count calendar days, others the last 24 hours on a rolling basis. “One email per day” then means two different things.
What no tool does for you is make the business decision on how your setup differentiates by signal:
- pauses whose length depends on the type of signal
- amplifiers that take effect after a click and slowly fade
- pauses across channels, staggered by cost
- pauses that depend on the case status in the service system
- rules for sales, which works in its own system
For many of these points, a data field, a filter in front of every flow and one person who maintains both are enough to get started.
Where to start
Start with the strongest signal. A field “explicit no: topic and date” and a filter that checks it before every send are enough to begin with. Build this for one topic, across all channels and all systems that select recipients. That puts principle 1 and principle 4 into practice on a small scale.
Complaints and purchases come next. Amplifiers come last, once the brakes work reliably.
Would you rather not go through your rule set alone, or are you looking for a second opinion? UNIT ONE goes through your contact rules with you along these five principles and prioritises which gap to close first. A good starting point is our Reactivation Audit.
That leaves one question for your next conversation with the CRM team: how many of your rules treat an explicit no differently from silence?
Frequently asked questions
What does contact pressure in CRM mean? It describes how many messages a customer receives across all channels in a given period, and how well they fit the customer’s situation. If it becomes too much, unsubscribes and spam complaints rise. In the end, this also hurts the delivery of the emails the customer wants.
Isn’t a cap per channel and week enough? As a start, it is better than no rule at all. But it only counts messages. It cannot see whether a customer has just bought something or complained. That takes rules that differentiate by signal.
How long should the pause after an explicit no last? Longer than after any other signal, and only for the declined topic. There is no universally correct value. The right duration depends on the topic and on the purchase cycle of your products.
Do invoices and other service messages count towards the cap? Better not. Otherwise an invoice pushes out the next newsletter, or a newsletter pushes out the invoice. Some blocks still apply to all messages, for example for an undeliverable address.
Where do we start? With the strongest signal. A field for the explicit no with topic and date, and a filter that checks it before every send, are enough at first. Build this for one topic and have every system that selects recipients read it.