Agency retainers are being unbundled

I’ve sent invoices with one line on them. “Digital marketing services - monthly retainer.” A number next to it. Nobody ever asked what was in it, which at the time I took as trust and now recognise as the beginning of the problem.

Gartner’s 2026 CMO Spend Survey, run across 401 marketing leaders between January and March 2026, found paid media at 31.4% of marketing budgets, a five-year high, funded by cuts to agencies. In-house labour rose from 21.9% to 24.5% over the same year. The money didn’t leave marketing. It moved inside the building.

A monthly retainer has always bundled 3 unrelated things: strategy, production and access. Production was the volume that made the number look justified, and production is the part that has just collapsed in price. So the bundle is coming apart, and what’s left underneath is the part that was always the actual value. Which is awkward, because it’s also the part most retainers were worst at pricing honestly.

The numbers, and where they come from

In-housing isn’t new. The Association of National Advertisers has tracked it for 15 years and found 82% of its members ran an in-house agency in 2023, up from 78% in 2018, 58% in 2013 and 42% in 2008. That’s a slow, steady line that predates any of this.

What’s new is the pace. Gartner’s 2025 survey of 402 marketing leaders found 39% planning to cut agency allocations that year. Advertiser Perceptions data cited by eMarketer put the six largest holding companies’ share of US ad spending at 29.6% in the first quarter of 2024, down from 44.6% in 2019.

There’s a widely quoted figure that 60% of senior US marketers spent less on agencies in 2025 because of AI, with 83% saying full automation of content production would remove most or all of their agency spend. It comes from a survey of 200-plus VP-and-above marketers published by Typeface in October 2025. Typeface sells AI content generation software, so that’s a vendor asking buyers whether they need the thing the vendor competes with. Directionally interesting. Not evidence.

The Australian picture is quieter and more useful. IBISWorld has digital advertising agencies here at $4.1 billion in 2025–26, growing 5.5% a year over 5 years, across 9,418 businesses that have grown at 9.7% a year since 2021. Revenue rising, agency count rising almost twice as fast. Do the arithmetic and average revenue per agency has been falling for 5 years. The market isn’t shrinking. It’s fragmenting, and the fee per relationship is going with it.

What agencies did to earn this

I ran a digital agency of my own for the better part of a decade, and a media production company of five after it. So this next part is a confession before it’s a criticism.

The retainer’s commercial logic was that production filled the hours. 10 social posts, a blog, a report, some ad variations. It was visible, it was measurable and it was the reason the number could be that size. Strategy was 2 hours a month, delivered in a meeting, and it was quietly the only irreplaceable thing on the invoice.

Then a set of habits grew up around defending that number. Hours that expired if you didn’t use them, which is a subscription with extra steps. A senior in the pitch and a junior on the file. Reports that measured activity because activity had gone up and results hadn’t. Scope defined loosely enough that the agency decided what counted. And the single-line invoice, which I sent, and which made the whole arrangement impossible to evaluate.

None of that was fraud. It was an industry that priced input because input was easy to count. When the cost of input fell through the floor, the pricing model had nothing left holding it up.

What to stop paying for and what to keep

If you run a business with real revenue and no marketing team of your own, this is the split I’d apply.

Stop paying a retainer for	Keep paying for
Volume content production	The positioning decision, made once, properly
Social posting and scheduling	Someone who tells you the campaign isn’t working
Standard report assembly	Media buying where the spend is large enough to matter
Design variations of an existing asset	Original creative: the idea, not the execution
Landing page variations	Technical work you can’t hire for
Hours that expire monthly	Accountability that survives a bad quarter

Three practical moves.

Ask for the invoice split. Strategy, production, media, access: 4 lines, 4 numbers. Any agency worth keeping can produce that in an afternoon. The reaction to being asked tells you as much as the answer.

Price access separately and don’t apologise for it. The most valuable thing I ever sold was being reachable by someone with a decision to make and no one to check it with. That has a price and it isn’t zero. It’s also not $8,000 a month with 12 social posts stapled to it.

Take production in-house on a trial, not a principle. Move one workstream, run it for a quarter and measure the real cost including your own team’s time. Most find production is cheaper in-house and that the strategic gap it exposes costs more than the saving. Better to find that out on one workstream than all of them.

The uncomfortable part, for the side of the table I sat on for a decade: an agency that can’t justify its fee line by line was never charging for the right thing.