Foreword
Earlier this month, we headed up to Manchester for Performance MCR, the North's practitioner-led performance marketing conference held at the Bridgewater Hall. Built by people who actually run paid media accounts day-to-day, it is the kind of event that skips the keynote fluff and gets straight to what is actually working in the real world.
We came away with pages of notes, a few shifted assumptions, and five ideas that we think deserve wider attention. Here is what stood out.
One of the sharpest reframes of the day was a simple but devastating observation: "Google scales products, not campaigns." The problem with dropping your entire product catalogue into a single automated campaign is that the platform will naturally identify a handful of bestsellers and funnel budget towards them. Everything else gets quietly ignored, no matter how strong the margin or how much potential it holds.
The proposed alternative is a Bid on Intent (BOI) framework, where every SKU is given a deliberate operational role. Is this product generating cash right now? Is it earmarked for clearance? Is it a slower-burn, higher-margin item that needs nurturing? Each role demands a different bidding approach and a different set of success metrics.
The mental shift here is significant. Rather than thinking of your Google Ads account as a campaign manager, think of it as a capital allocation tool. Every pound of budget is a decision about which products to back and why.
Automated scripts that pause or activate products based on stock levels or performance thresholds are nothing new. What is less common is building genuine intelligence into how those scripts define "performance."
Most standard scripts treat all products as equivalent, regardless of how they actually sell. A seasonal item with a natural 90-day selling window gets judged against the same benchmarks as a perennial bestseller, which is a category error.
The more sophisticated approach involves adding a rolling 30-day "Days to Sell" velocity metric directly into the script logic. This allows the system to dynamically push and pull products based on where they genuinely sit in their natural lifecycle. A product with accelerating velocity gets more investment. One that is slowing down, or behaving seasonally, gets treated accordingly, automatically and in real time.
The takeaway for marketers is that the script is only as useful as the data layer underneath it. If you are treating every product the same, you are leaving a significant amount of decision-making quality on the table.
This one will sting for anyone who has spent time in a Meta or Google interface recently. Platform automation has quietly pushed brands towards a production-line model: create more variants, feed the algorithm, let it optimise.
The problem is that when all of your creative variants are built on the same underlying template, the same message, the same visual approach, the algorithm serves them all to broadly the same audience cluster. Volume increases, but genuine reach does not.
What actually unlocks new audience signals is creative difference, not creative volume. The concept introduced here is a Creative Diversification Scoring Matrix: an audit framework for your existing asset library that scores how genuinely diverse your creative approach really is.
A score of 5, the highest, reflects a library with three or more distinct message types fully diversified across audience segments and emotional hooks, four levels of brand expression in the creative approach, and the full range of core media formats including user-generated content, across all Advantage+ placements.
A score of 1 or 2 looks like the opposite: two message types, limited brand expression, a narrow format range, and somewhere between six and ten placements.
The practical implication is clear. Before you scale production, audit for genuine difference. If your creative library scores low, adding more of the same will not expand your reach; it will just entrench the audience ceiling you already have.
Creative hooks are not standalone units. They exist at the start of a path, and what happens next on that path determines whether the hook was wasted or not.
This is particularly acute for products with a higher price point than direct competitors. If your creative does its job, it creates a cognitive shift in the viewer. It reframes the value equation and gets someone genuinely interested. But if that person lands on a standard, transactional product page, the page immediately disrupts that shift. There is nothing there to cement the new perspective. They bounce, and the budget spent on that brilliant hook has done nothing.
The argument here is that marketers need to stop A/B testing minor variations of ad hooks and start testing the entire user journey. One practical approach: send cold traffic from a bold, attention-grabbing creative to an educational landing page that reinforces the value case, rather than to a direct product page. Then retarget those warmed-up users separately, converting them from a qualified pool rather than asking them to make a decision they are not yet ready for.
It is a longer conversion path, but it is one that respects both the creative investment and the cognitive work required to sell a premium product to a cold audience.
The final theme was one of the most timely: how to work with highly automated campaign types like AI Max, which bundles Broad Match, Performance Max, and Dynamic Search Ads together into a single machine.
The honest assessment: left entirely on autopilot, these tools can double your impression volume almost immediately. They can also bleed your click-through rates and accumulate significant low-intent waste just as quickly.
The recommended posture is to treat AI automation as an aggressive experiment that requires clear human-set boundaries, not a set-and-forget layer. In practice, that means three things.
First, only deploy it once your account is generating enough conversion volume to give the platform real signal to work with. The suggested threshold is at least 100 conversions per month on an unrestricted budget.
Second, enforce rigid text customisations. That includes excluding competitor brand names, blocking low-quality intent modifiers like "free" or "cheap," and applying regional formatting where relevant.
Third, use Seasonality Adjustments actively. When target CPAs start rising under budget pressure, manually force the platform's bidding downward rather than waiting for the algorithm to self-correct, because it often will not do so quickly enough.
The broader point here is one that cuts across all five takeaways: automation does not replace strategic thinking, it amplifies whatever decisions are already embedded in the account. The marketers who get the best results from these tools are not the ones who hand over the most control; they are the ones who define the clearest constraints and then let the machine work within them.
Performance MCR is an annual conference held at the Bridgewater Hall in Manchester. It is built for practitioners running paid media accounts and is led by people who are still doing the job themselves. If any of the ideas above resonate with how you are thinking about your own paid media approach, we would love to talk it through, so get in touch.
