Four Newsletters Capped at 3. Your Reader Saw It 11 Times.

Four newsletters in the same category, each capping your campaign at three exposures per subscriber per week. Every publisher is compliant. Your insertion order said three and every party honoured it.

A subscriber to all four sees your advertisement eleven times.

The campaign report will show an average frequency close to three, because it is calculated per publisher and averaged. That number is accurate and it describes something other than what happened to the reader.

This guide walks through the week exposure by exposure, explains why no single publisher can fix it, covers the identity mechanism that actually resolves it, and gives both sides the questions to ask before the next flight.

The week, day by day

A reader subscribes to four newsletters covering personal finance. This is not unusual behaviour. It is what an engaged reader in a subject does, and the same pattern holds in technology, marketing, health, and every vertical where several good publications compete for one audience.

An advertiser buys across all four, which is also normal, because those four are the obvious buy for that audience.

Each publisher caps at three exposures per subscriber per week.

• Monday. Newsletter A sends. The reader opens. Publisher A's counter for this subscriber: 1.
• Tuesday. Newsletter B sends. Same reader, same ad. Publisher B's counter: 1. B has no knowledge that Monday happened.
• Wednesday. Newsletter C sends, counter 1. Newsletter A sends a mid-week edition, counter 2.
• Thursday. Newsletter D sends, counter 1. Newsletter B sends again, counter 2.
• Friday. A sends again, counter 3, cap reached, and A stops serving. C sends again, counter 2. D sends again, counter 2.
• Saturday. B's weekend edition, counter 3, cap reached.

Total exposures for that reader across the week: eleven.

No publisher exceeded three. The reported campaign frequency, averaged across four publishers each showing between two and three, lands somewhere near two and a half.

Why the reported number and the experienced number diverge

The campaign report is not lying. It measures per-publisher exposure, and every value in it is correct.

It simply does not measure the thing that determines the campaign's effect, which is total exposure per human being.

Those two numbers diverge in direct proportion to subscriber overlap between the publishers on the buy. And here is the property that makes this genuinely awkward: subscriber overlap is highest exactly where a campaign is best targeted.

Good targeting means buying the publications your audience actually reads. Your audience reads several of them. The better your media plan, the more overlap it contains, and the worse this gets.

Industry guidance on frequency generally lands in the range of two to five exposures per user per week, with research cited around a threshold near six beyond which return on investment degrades meaningfully. Eleven is comfortably outside any of those ranges, and it was produced by four parties each observing a cap of three.

What eleven exposures costs the advertiser

Frequency has a shape. Early exposures build recognition and consideration. A middle range continues adding value at a declining rate. Beyond some point, additional exposure stops helping and begins working against the advertiser.

Where that turn sits varies by category, creative, and purchase cycle, and I am not going to invent a precise number for it. What is not disputed is that the curve turns, and that eleven exposures of an identical creative to one person inside seven days sits past the turn for essentially any campaign.

Past that point the advertiser is paying for exposures that reduce brand favourability. Not neutral impressions. Negative ones, purchased at full rate.

There is a second cost that is larger and less obvious. Unique reach comes in far under plan, because budget that should have reached new households went into re-hitting the same one. The advertiser paid for reach and received frequency.

Then the response makes it worse. A mid-flight report showing reach under plan usually triggers one of two actions: increase budget, or add more publications in the same category. Increasing budget buys more impressions against the same overlapping audience. Adding publications in the same category adds more overlapping subscriber bases, because publications in a category share readers by definition. The reader who was on four lists is plausibly on six.

The overlap arithmetic, and what it does to your reach number

The day-by-day walkthrough shows what happens to one reader. The campaign-level version is where the budget actually goes.

Take a buy across four newsletters with 30,000, 25,000, 20,000 and 15,000 subscribers. Summed, that is 90,000, and 90,000 is the number that appears on the media plan as addressable audience.

Now assume meaningful overlap, as you should on a tightly targeted category buy. Say unique subscribers across all four come to 58,000. That means 32,000 of the summed total, roughly 36%, is the same people counted more than once.

The campaign delivers three exposures per subscriber per publisher, so 270,000 impressions in total.

Divide by actual unique reach: 270,000 ÷ 58,000 = an average frequency of 4.66, against a stated cap of three.

And that is the average. The reader on one list saw three. The reader on all four saw eleven. The distribution is wide and the reported average conceals both ends of it.

Two numbers matter here and only one appears in most reporting. Impressions delivered in full, so the campaign looks complete. Unique reach came in 36% below the planned addressable audience, which appears nowhere unless someone deduplicates.

The useful diagnostic: divide total impressions by unique subscribers reached. If that figure is materially above your stated cap, overlap is consuming your budget. If your reporting cannot produce unique subscribers reached across publishers, you cannot run the check, and that absence is the finding. Modelling this against your own numbers is straightforward with the newsletter ad revenue calculator.

Why the publisher pays for it too

Publishers frequently regard this as an advertiser problem, and there is a defensible version of that view, since each publisher honoured the cap they agreed to.

The consequence lands on them regardless, in two ways.

The reader blames the newsletter. Someone who saw the same advertisement eleven times does not attribute that to the advertiser's media plan. They attribute it to the publications that delivered it. Over-frequency irritation routes into the unsubscribe-or-complain decision, and complaint rate is one of the few things mailbox providers actively grade you on. The mechanics of why that matters commercially are covered in deliverability as an ad revenue problem.

The advertiser does not renew. Evaluating a campaign that under-delivered on reach and produced fatigue, the advertiser concludes the channel underperformed. They will not diagnose cross-publisher overlap, because nothing in their reporting surfaces it. They move budget elsewhere, and no publisher in the chain learns why.

Both sides lose, and no party did anything wrong.

Why no single participant can fix it

This is the part that explains why the problem persisted quietly for years rather than being an oversight.

Publisher A cannot cap against exposures it cannot see. It has no visibility into what B, C, or D served.

The advertiser cannot enforce a cap across publishers who each hold their own subscriber records with no mechanism to compare them.

No publisher will share their subscriber list with a competitor to make frequency capping work, and nobody should expect them to.

The underlying obstacle is identity. These are four independent publishers with four independent subscriber databases and no shared identifier. Without something that lets them recognise the same human across four lists, and without exposing the lists, a cap has nothing to count against.

The workarounds publishers reach for do not address this. Asking the advertiser to lower their stated cap does not help, because the cap was never the binding constraint: four publishers capping at two produces eight, which is still past the useful range. Staggering send days does not help either, since the overlap is in the audience rather than in the calendar.

How hashed identifiers solve it without sharing lists

The mechanism is a shared frequency layer sitting above the individual publishers, counting exposures per subscriber across all of them, using a privacy-preserving identifier rather than the email address itself.

In practice this means each publisher applies the same hash function — typically SHA-256 — to a normalised version of the subscriber's email address before it leaves their system. The exchange receives the hash.

What the exchange can do with that: recognise that the hash it saw on Monday for Publisher A is the same hash it is seeing on Tuesday for Publisher B, and refuse to serve once the cap is reached.

What it cannot do: reverse the hash to recover an address, reconstruct any publisher's subscriber list, or tell Publisher A anything about Publisher B's audience. Each publisher sees only their own subscribers; the exchange sees only opaque identifiers and a count.

Normalisation matters more than people expect. The same person can appear as different strings across four lists: differing case, Gmail dot variations, plus-addressing. Hashing without normalising first produces four different hashes for one human and the cap silently fails to bind. The implementation detail is covered in audience targeting with email hashes.

This is the same problem connected television had to solve at the household level, and it took that industry several years to accept that per-publisher capping was not capping at all. Email advertising is at the point in that arc where the problem is understood by the people running the infrastructure and largely invisible to the people buying the media.

Where the buying route changes the answer

Frequency behaves differently depending on how the inventory was purchased, and the difference is usually invisible during planning.

Direct-sold across several publishers is the worst case. No intermediary counts anything. Each publisher enforces its own cap against its own records, and nothing observes the total. An advertiser running four independent sponsorships has four independently correct caps and no cap at all in the sense that matters.

A single publisher with multiple newsletters is the best case, and it is underrated. A publisher operating several titles can enforce a genuine cap across all of them, because the subscriber records sit in one system. If you are buying a portfolio from one operator, ask whether the cap spans the titles — frequently it does, and it is rarely mentioned as a selling point.

Exchange-mediated buying resolves it where the exchange holds a shared identifier across participating publishers, which is the mechanism described above. The trade-offs between routes are covered in sponsorships versus programmatic.

A mixed buy is the trap. Part of the budget through an exchange with real capping, part direct-sold outside it. The exchange enforces correctly across its own inventory and has no visibility into the direct-sold placements. A household can be capped correctly within one half of your buy and hit repeatedly by the other.

This is the most common structure among advertisers who take frequency seriously, which is a slightly grim irony. The mitigation is to route the direct-sold placements through the same identity layer where the publisher will accept it, or failing that to stagger the direct flights against the programmatic ones rather than running both concurrently. Campaign structure options are in the advertiser portal.

What a publisher can measure without any shared infrastructure

Publishers usually assume overlap is unknowable from their side. It is partly measurable, and the partial version is useful.

Ask advertisers what else they are running. Not a technical solution, and it is the cheapest available information. An advertiser who names the other three newsletters on the buy has told you your overlap exposure for that campaign, and most will answer because it costs them nothing.

Watch your unsubscribe rate against ad-heavy sends. If unsubscribes spike on issues carrying a specific advertiser, and that advertiser is running broadly in your category, over-frequency elsewhere is a plausible cause even though the exposure did not happen on your list. Your subscribers are reacting to a total you cannot see.

Track click-through decay across a flight. A campaign whose click rate declines steeply across successive sends on your newsletter alone is showing fatigue. If the decline is steeper than your historical pattern for comparable creative, the reader is likely arriving at your send already saturated. Per-send campaign trends are visible in reporting.

Compare performance for the same advertiser across flights. If an advertiser performs well in a quarter when they are running narrowly and poorly when running broadly across your category, that pattern is overlap, and it is worth raising with them directly. It reframes a disappointing result as a media-planning issue rather than an inventory-quality one, which is a considerably better conversation to have.

What frequency should actually be set to

Once a cap genuinely binds across the buy, the number becomes a real decision rather than a formality.

Start lower than you think. A cap of three that actually binds across four publishers is materially more exposure than a cap of three that binds within each. If you are moving from per-publisher to cross-publisher enforcement, your effective frequency is about to fall substantially at the same nominal setting, which is the point.

Set it against the purchase cycle. A considered purchase with a long evaluation period tolerates and benefits from more exposures spread over more weeks. An impulse or low-consideration product needs fewer, closer together. A single number applied across all campaigns is a default rather than a decision.

Vary the creative before you raise the cap. Fatigue is partly a function of repetition of the same creative rather than of exposure count alone. Sequential creative across exposures behaves differently from the same unit eleven times — which is the mechanism behind ad journeys and covered in sequential newsletter campaigns.

Watch reach and frequency together, never separately. A campaign hitting its reach target with frequency at the cap is working. Reach under target with frequency at the cap means overlap is eating your budget, and the fix is broader targeting or different publishers rather than more money.

Questions to ask before the next flight

If you are buying: is frequency enforced across the whole buy, or per publisher and then averaged in the report? Most sellers will answer honestly because most have not been asked. The answer tells you whether the number on your insertion order is a policy or a decoration.

Then: can you report unique subscribers reached, deduplicated across publishers, rather than total impressions? A seller who can produce a deduplicated reach figure has the identity infrastructure to enforce a cap. One who cannot, does not, whatever the insertion order says.

If you are selling: can you tell an advertiser what share of your subscribers also appear on other publishers in their buy? You do not need to know which publishers — only that the overlap exists and its rough magnitude. That single figure changes how a sophisticated buyer values your inventory, usually upward, because it demonstrates you understand what you are selling.

And: when a campaign under-delivers on reach, do you have the data to show whether overlap caused it? A publisher who can explain a disappointing result keeps the account. One who cannot loses it to a cause nobody identified. This belongs alongside the other diligence questions in the advertiser vetting checklist.

Frequently asked questions

Can I just ask publishers to coordinate caps between themselves?

In principle yes, in practice almost never. Coordination requires each publisher to disclose which subscribers they served, which is the one thing none of them will do for a competitor. Even where publishers are willing, there is no common identifier to coordinate against, so the conversation stalls at implementation rather than at willingness. This is precisely the gap a shared hashed-identifier layer fills — it lets the counting happen without anyone disclosing anything.

Why does per-publisher frequency capping not work?

Because each publisher can only count exposures it served. A subscriber on four newsletters in the same category can receive a campaign from all four, each honouring its own cap, producing a total far above what the advertiser specified. Caps only bind when something counts across all publishers in the buy against a shared subscriber identifier.

Does sharing hashed emails expose my subscriber list?

No. A SHA-256 hash of a normalised email address is a one-way transformation — it cannot be reversed to recover the address. An exchange holding hashes can recognise that two publishers served the same person without learning who that person is, and cannot enumerate or reconstruct any publisher's list. Each publisher continues to see only their own subscribers.

What frequency cap should I set for newsletter advertising?

Common industry guidance sits between two and five exposures per subscriber per week, with evidence that returns degrade meaningfully past roughly six. The more important point is that the number only means anything if it binds across every publisher in the buy. A cap of three enforced per publisher across four newsletters is not a cap of three.

How do I know if overlap is affecting my campaign?

Compare unique reach against plan. If impressions delivered in full while unique subscribers reached came in well under target, overlap is the most likely explanation, particularly on a tightly targeted buy across publications in one category. If your reporting cannot produce a deduplicated reach figure at all, that absence is itself the finding.

Will lowering my cap fix the problem?

Not on its own. Four publishers capping at two produces eight exposures rather than eleven, which is still past the useful range. Lowering caps reduces the symptom without addressing the mechanism, and it costs you delivery on publishers where the reader is not overlapping. Cross-publisher enforcement is the fix; lower caps are a partial mitigation.

Is overlap worse in niche categories or broad ones?

Worse in niche categories, which is counterintuitive until you consider where readers come from. A narrow professional subject has a limited pool of interested people and a handful of publications serving them, so a genuinely engaged reader is likely subscribed to several. A broad consumer category has a far larger pool spread across many more publications, so any two lists overlap less as a proportion.

This means the buys most likely to suffer are the tightly targeted ones a sophisticated advertiser would choose deliberately. Precision targeting and high overlap are the same property viewed from two sides.

Does this apply to direct-sold sponsorships too?

Yes, and it is harder to address there. A brand running direct sponsorships with several newsletters independently has no intermediary counting anything, so overlap is entirely unmanaged. The advertiser can mitigate by staggering flights across publishers rather than running them concurrently, which spreads exposures over more weeks without requiring shared identity. It is a blunter instrument than a shared frequency layer and considerably better than nothing.