Solved: Ad Group/Audience Structure for Video Action Campaign

Hi there,

Thanks for reaching out! Happy to give you some thoughts on how to structure your Google Video Action campaigns. It's a common question, and honestly, the way most people think about it is a bit outdated and can really hold back performance.

You're right in the middle of a classic dilemma: granular control versus data consolidation. The short answer is that for modern, conversion-focused campaigns like this, you almost always want to lean heavily into consolidation. Let’s get into why that is and how you can do it properly without losing all control.

We'll need to look at how Google's AI really works...

First off, you have to forget a lot of the old rules about Google Ads. A few years ago, the best practice was absolutely to have hyper-segmented ad groups. One audience per ad group, sometimes even one keyword per ad group in search. This gave the advertiser maximum manual control over bids, ads, and budget. But the platform has changed dramatically. Campaigns like Video Action (and now Performance Max) are built on machine learning. Their entire purpose is to take your inputs—your budget, your conversion goal, your creatives, and your audience signals—and find the cheapest conversions for you.

The most important resource for the algorithm is data. Lots and lots of data points, or 'signals' as Google calls them. When you split your 10 audiences across 10 different ad groups, you're essentialy telling Google to run 10 seperate, tiny campaigns. Each ad group gets a small slice of the budget and a trickle of data. The algorithm in Ad Group A has no idea what the algorithm in Ad Group B is learning. It's incredibly inefficient. Each ad group will likely struggle to get out of the 'learning phase', and you'll never give the system enough data to properly identify what a high-converting user actually looks like across your entire potential audience.

Think of it like this: you've hired a world-class chef (the Google algorithm) and you want them to cook the best meal possible (find you conversions). The old way is like giving the chef ten tiny saucepans, each with only one ingredient, and telling them to make a gourmet dish. It's not going to work well. The modern approach is to give the chef a single, massive pot and all ten ingredients at once. The chef has the skill and experience to know which ingredients to use more of, which to use less of, and how to combine them for the best result. That's what you're doing when you consolidate your audiences.

To really get your head around this, it helps to visualise the flow of data in both scenarios.

I'd say you should combine your audiences...

So, my direct advice is this: put all 10 of your related in-market and custom intent audiences into a single ad group. Don't seperate them. By doing this, you're giving Google's AI the clearest possible picture of the types of people you want to reach. The system is incredibly good at figuring out which nuances within those audiences are most likely to convert. It will automatically start to shift more of the budget towards the specific segments that are delivering results, without you needing to do anything manually.

Your intuition that "more data in one ad group might be better" is spot on. It's not just a bit better; it's fundamentally how these campaigns are designed to succeed. Let’s imagine you have a £100 daily budget. In the fragmented model, each of your 10 ad groups gets £10 per day. At a £1 Cost Per Click, that’s just 10 clicks per ad group. How long do you think it will take to get even one conversion, let alone the 20-30 needed for the algorithm to learn anything meaningful? It could take weeks, or it might never happen. In the consolidated model, your single ad group gets the full £100 per day. It gets 100 clicks. It might get 5-10 conversions on day one. The system now has a wealth of data to analyse and can start optimising immediately.

This isn't just theory; it's something we've seen deliver powerful results. I remember a campaign we worked on for a medical job-matching software client. They were running campaigns on Google Ads, but their cost per user acquisition was hovering around a costly £100. Their structure was fragmented, which starved the algorithm of the data it needed to optimise effectively. After we restructured their campaigns, consolidating their targeting and letting the algorithm work with a much larger dataset, we saw a dramatic change. We were able to reduce their cost per user acquisition all the way down to just £7. The core change was trusting the machine and giving it the right structure to succeed.

You probably should focus on creative, not just structure...

Here’s the other part of the equation that most advertisers miss. They get so bogged down in audience settings and bidding strategies that they forget the single most important element of a video campaign: the video itself. Your campaign structure is the delivery vehicle, but your creative is the message. You could have the most perfectly structured campaign in the world, but if your video ad is boring, unclear, or doesn't resonate with the audience, it will fail.

Instead of spending your time building 10 different ad groups, you should be using that time to create 2-3 different video ads to run within your single, consolidated ad group. This is how you really move the needle on performance. The algorithm will not only figure out which *people* are most likely to convert, but it will also figure out which *message* is most effective at converting them. It will automatically serve the best performing ad more often, maximising your results.

What should you test? Dont just make tiny changes like a different background colour. Test fundamentally different approaches and hooks. Here are a few frameworks we use:

• Problem-Agitate-Solve (PAS): Start by calling out a specific pain point your audience experiences. Agitate that pain by explaining the negative consequences. Then, present your product as the clear solution. This is powerful for audiences who are problem-aware.
• Before-After-Bridge (BAB): Show the desirable "after" state. Paint a picture of what life is like once their problem is solved. Then, show the undesirable "before" state they're currently in. Finally, present your product as the bridge that gets them from before to after.
• UGC/Testimonial Style: Use a customer testimonial or a simple, phone-shot video that looks like user-generated content. This builds authenticity and trust, and often cuts through the noise of slick, professional ads. We've seen SaaS clients get amazing results with simple screen recordings narrated by a founder.

By testing these different angles, you give the algorithm options. It might discover that your in-market audience responds best to the direct PAS approach, while a custom intent audience (who might be further down the funnel) converts better from the BAB video. With a consolidated ad group, the system can make these connections automatically and optimise in real-time.

You'll need to manage performance the right way...

Now, I know what you're thinking. "If I put everyone in one ad group, how do I maintain any control? What if one audience is wasting all my money?" This is a valid concern, and it's where modern performance management comes in. You don't need seperate ad groups to have control.

Your control comes from analysis, not seperation. Inside your single ad group, navigate to the "Audiences" tab on the left-hand menu. This is your command centre. Here, Google will break down the performance of every single audience segment you added. You'll see impressions, clicks, cost, conversions, and cost per conversion for each one. Let the campaign run for a week or two until you have statistically significant data (e.g., a few hundred clicks or a decent number of conversions for each segment).

Now you can make informed decisions. Is one of your in-market audiences getting a lot of clicks but zero conversions, driving up your overall CPA? Select that audience and exclude it from the ad group. That's it. You've just trimmed the fat without disrupting the learning of the wider ad group. This is a much more surgical and effective way to optimise than simply pausing an entire ad group that might have had some potential. You are pruning the branches, not cutting down the tree.

To help you get a feel for the numbers, I've built a simple interactive calculator. You can adjust your expected spend, click cost, and conversion rate to see how these metrics influence your overall lead volume and cost. It's a good way to set realistic expectations before you launch.

I've detailed my main recommendations for you below:

To pull all this together, here is a clear summary of the strategy I'd recommend you implement. This approach moves away from unnecessary complexity and focuses on what actually drives conversions in today's Google Ads environment: strong creative and intelligent use of the algorithm.

Following this structure should set you on a much better path. It simplifies your management workload and, more importantly, aligns your campaign with how Google's systems are designed to work, which usually leads to better and more scalable results.

Of course, this is just the foundational strategy. The real work comes in the ongoing testing of creative, analysing the audience data, optimising the landing page, and knowing when to scale. It can be a lot to manage, and small mistakes can be costly. If you want to make sure you're getting the absolute most out of your budget and want to accelerate your results, it can often help to have an expert take a look.

We offer a completely free, no-obligation strategy consultation where we can take a look at your specific situation and give you a more detailed plan of action. Feel free to get in touch if that sounds helpful.

Hope this helps!

Regards,

Team @ Lukas Holschuh