When platforms take control, B2B pays the price

As Google e and Meta automate more of their advert­ising platforms, B2B marketers face a new challenge to ensure campaigns optimise for business value, not simply more conversions. 

At Comprend, we work with B2B organisations navigating increasingly complex digital marketing environments. As advert­ising platforms take greater control over campaign delivery, the challenge is shifting from managing individual settings to ensuring automation delivers meaningful business outcomes.

Last September, Google upgraded eligible Search campaigns to AI Max. The migration happened in place, campaign by campaign across the month, and once the window opened there was no opt-out or rollback. Most accounts received no warning email.

In August, Meta started removing placement exclusions from ad sets. The replacement is a value rule that adjusts bids for a placement between -90% and +1,000%, allowing advert­isers to make a placement more expensive without excluding it completely.

We're seeing the same direction of travel across two of the biggest advert­ising platforms within a short period. Campaign control as we know it is changing.

Meta's numbers and what the average hides

Meta has published its reasoning behind this approach. Its documentation puts the average cost per action for ad sets using Advantage+ placements at 11.7% below ad sets with manual placements.

That average covers advert­isers of every skill level, including accounts that switched off placements out of habit without checking whether it helped. Against that baseline, handing delivery to the model can be an upgrade.

Google's reasoning on AI Max follows a similar logic. For many accounts, it works: the system finds conversions that tighter keyword control may have missed.

The automation works. But what it works towards depends on the signals you give it, and that's where B2B diverges from other advert­isers.

What these models need and what B2B can't always supply

Every automated delivery system relies on conversion signals. The more signals available, the more effectively it can optimise.

A consumer account might feed the model thousands of purchases a month, each a clear outcome with a value attached. A B2B account in a specialised industrial category might generate 40 qualified leads in a good quarter.

At that volume, the system struggles to distinguish between a procurement director at a target account and a student filling in a form for coursework. Both count as conversions, even though one could represent a six-figure sales opportunity.

In other words, the model optimises towards the signals it can read. It finds the cheapest conversions available, on the placements where cheap conversions occur, from audiences most willing to provide an email address.

Cost per lead falls, and your dashboard looks healthier than it did last quarter.

Meanwhile, your sales team no longer recognises the company names coming through. Two quarters of budget may have reached an audience that was never going to buy. And the campaign setting that changed in September is investigated last, because the reporting never flagged a problem.

Oversight is the job now

If platforms have taken the operational controls, what's left is strategic – and it's arguably more important than the work it replaced.

Start with the signals. A system optimising towards form submissions will deliver form submissions. Feed it signals that reflect business value instead: offline conversion imports, value-based signals from your CRM and lead scores passed back to the platform.

Many B2B accounts still haven't done this, yet it can change what the algorithm prioritises more than any remaining campaign setting.

This is also where tech-enabled marketing and communications comes into its own, because the strategy only works if the data infrastructure behind it does.

Move hard exclusions to the controls that remain. Meta's account-level placement controls, under Advert­ising Settings, remain a genuine block and apply across campaigns. Google carried brand inclusions and exclusions through the AI Max migration unchanged.

Both are broader than the controls they replaced. Decisions about where a brand should and shouldn't appear therefore need to be made at account level, with a clear view of the brand's priorities. That's a strategic decision, rather than a checkbox.

Measure beyond the platform. If the automation reports on conversions and you only measure conversions, you're marking the model's homework with its own answer sheet.

Pipeline contribution, opportunity quality and deal velocity provide a clearer picture of business impact. For organisations with longer sales cycles, these are the measures that have always mattered.

Where does B2B paid media go next?

Dynamic Search Ads are expected to follow in February 2027, with more manual controls likely to disappear over time. Arguing against automation is unlikely to be a productive use of anyone's time.

The job has changed shape.

Running campaigns used to mean choosing placements, matching keywords and writing ads. Platforms now handle much of this to a reasonable standard.

What they don't know is what your business considers a good customer, which audiences you'd rather not reach or what a qualified lead is worth to your sales team.

That knowledge used to sit in campaign settings where everyone could see it. Now it sits in signals, campaign structure and measurement, where building it takes longer and neglecting it costs more.

Over the next 18 months, the B2B advert­isers that pull ahead will be those that deliberately define what their automation is optimising towards.

If you run B2B paid media and haven't reviewed what changed in your accounts in September, that's a good place to start. Let's talk.