Hi there,
Thanks for reaching out!
Happy to give you some initial thoughts on your PMax situation. It's a really common question and tbh, there's a lot of dodgy advice out there about it. The short answer is that you're right to be cautious, but you might be worrying about the wrong thing. It's not the budget setting itself that'll trip you up, but how you manage the bidding strategy and signals around your seasonal shifts. Let's get into it.
TLDR;
- Your budget ceiling doesn't 'confuse' the algorithm; it's just a cap. The real driver is your tCPA target and the conversion data it's learned from.
- The main danger with seasonality is the algorithm using peak-season data to bid during the slow season, causing it to overspend or stop spending entirely.
- Don't just let the campaign run on autopilot. You need to proactively manage your tCPA target based on the season. This often means increasing your tCPA target during slow times (sounds weird, but it's correct).
- Scaling budget should always be slow and steady – no more than 20% every few days – to avoid triggering the dreaded 'learning phase' all over again.
- This letter includes an interactive calculator to help you figure out your true Lifetime Value (LTV), which is essential for setting realistic CPA targets, and a flowchart to visualise how PMax actually thinks about your budget.
Let's bust a myth: PMax doesn't get 'messed up' by a high budget
First off, let's clear up this big misconception. Google's algorithm, especially in a black box like Performance Max, isn't a person. It doesn't get confused or offended if you give it a £500/day budget and it only spends £200. It doesn't sit there thinking, "Hmm, they gave me all this money, I MUST spend it!".
Think of it like this: your budget is the *maximum* you're willing to spend in a day. It's a ceiling, a safety net. Your tCPA (Target Cost Per Acquisition) target is the actual *instruction*. You're telling Google, "Go find me as many conversions as you can, but I'm not willing to pay more than, say, £25 on average for each one."
So, on a busy day in your peak season, there might be loads of people searching for your stuff who look like they're ready to buy. The algorithm sees hundreds of opportunities where it predicts it can get a conversion for around £25. It will seize those opportunities and spend right up to your £500 budget if it can.
On a slow day in the off-season, the opposite happens. There are far fewer people searching, and those who are might be less likely to convert. The algorithm looks for opportunities to hit your £25 tCPA target, finds only a handful, and so it only spends £200. It's doing exactly what you told it to do. The unspent £300 isn't a negative signal; it's a sign the system is respecting your primary instruction (the tCPA) over the secondary permission (the budget).
Worrying about the budget setting is a red herring. The real issue, and the thing that catches most ecom businesses out, is managing the system's *expectations* as you move between seasons.
Your Daily Budget
e.g., £500/day (A Ceiling)
The PMax Algorithm
Is there an auction where I can likely get a conversion for ~£25 (Your tCPA)?
Result
If YES, bid. If NO, don't bid. Spend is based on opportunity, not budget.
You'll need to manage the algorithm's memory, not its budget
The real danger with seasonality is signal lag. PMax bases its predictions on past performance, typically looking at a conversion window of 7, 30, or even 90 days. So, imagine your peak season ends on January 31st. On February 1st, demand drops off a cliff. But for the next few weeks, your PMax campaign is still looking at all that juicy conversion data from January. It still *thinks* conversions are easy and cheap to get.
This creates two potential problems:
- It overbids and wastes money: The algo enters auctions thinking a user is a 10/10 prospect based on January's data, but in reality, they're a 3/10 February browser. It bids high, doesn't get the conversion, and your CPA shoots up.
- It panics and stops spending: After a few days of overbidding and failing, the system might overcorrect. It concludes, "Wow, it's impossible to hit my £25 tCPA target anymore," and it dramatically reduces your ad delivery, sometimes to almost nothing. You go from a flood of traffic to a trickle.
This is where most people get it wrong. They see the spend drop and think, "I need to raise the budget!" or "I need to lower my CPA target to get cheaper conversions!". Both are often the wrong move. You don't have a budget problem; you have a data problem. You need to give the system new instructions that reflect the new, harsher reality of the off-season.
I'd say you need a proactive seasonal strategy
You can't just set and forget a PMax campaign for a seasonal business. You have to be a bit more hands-on. Here's what we typically do for our e-commerce clients in a similar boat.
1. Adjust Your tCPA Target for the Season
This is the most important lever you have. It sounds completely backwards, but as you enter your slow season, you often need to *increase* your tCPA target. If your peak season tCPA was £25, you might need to tell Google you're now willing to pay £35 or £40 in the off-season. Why? Because you're acknowledging reality. Conversions are harder to come by, so you have to be willing to pay more for the few that are available. This tells the algorithm, "It's okay, I know things are tougher now. I'm giving you more breathing room to find those rare customers." It stops the system from panicking and shutting down your ads.
Conversely, as you approach your busy season, you should do the opposite. A few weeks before the rush, start gradually *lowering* your tCPA target. This tightens the leash and tells the algorithm to be more aggressive and efficient as more and more high-intent customers enter the market.
2. Use Seasonality Adjustments
This is a tool inside Google Ads that not many people use properly. It’s designed for exactly your situation. You can tell Google, "Hey, between February 1st and April 30th, I expect my conversion rate to be 30% lower than usual." The system will then factor this information into its bidding, making it much more intelligent about how it adjusts for the slowdown. It helps prevent that 'panic' we talked about. It's a way of giving the algo a heads-up so it's not caught by surprise.
3. Refresh Your Signals and Assets
Since PMax relies so heavily on the signals you give it, you need to make sure they're relevant.
- -> Audience Signals: Do you have different customer lists for peak-season buyers vs. off-season buyers? You could create lookalike audiences from these seperate groups and feed them into the campaign as signals to help it find the right kind of person for that time of year.
- -> Asset Groups: You should absolutely have different asset groups for different seasons. Your ad copy, images, and videos for your peak season should have a sense of urgency and excitement. Your off-season assets might focus more on value, benefits, or planning for the next peak season. A "Get Ready for Summer" campaign in Spring is a classic example. Don't just run the same ads all year round; it's a huge missed oppertunity.
You probably should look at your product feed more closely
You mentioned you're running a feed-only PMax campaign. This is great for simplifying things, but it means your entire campaign's success rests on the quality of that feed. It's your single most important asset, and most businesses neglect it.
Optimising your feed isn't just a one-time job; it's an ongoing process. Here are a few things that can make a massive differance:
- Product Titles: Are they structured for search? A good formula is `Brand + Product Type + Key Attribute (Colour, Size, etc.)`. Don't just put "Blue T-Shirt". It should be "YourBrand Men's Crew Neck T-Shirt - Royal Blue - Cotton". The more detail, the better Google can match it to relevant searches.
- Custom Labels: This is your secret weapon for seasonality. You can use custom labels in your Merchant Centre feed to tag products. For example: `custom_label_0 = "summer_bestseller"`, `custom_label_1 = "winter_clearance"`, `custom_label_2 = "high_margin"`. You can then use these labels to subdivide your PMax campaign or, at the very least, get much better reporting on what's actually selling when.
- Image Quality: It's shocking how many businesses use poor-quality images. Your main product image should be on a clean white background. But for your other images, show the product in use. Use lifestyle shots. Show people enjoying the product. We've seen conversion rates jump just by improving the quality and variety of product photos. One campaign we managed for a women's apparel brand saw a 691% return, and a big part of that was relentlessly testing and optimising their product imagery in the feed.
A well-maintained feed is the foundation. Without it, you're building your entire advertising strategy on sand. It doesn't matter how clever your bidding is if the product information is weak.
We'll need to look at if tCPA is even right for you
One last point. You're using tCPA, which optimises for the *number* of conversions at a certain cost. For an e-commerce store, this is only a good idea if all your products have a very similar price and profit margin. If you sell some items for £20 and others for £200, tCPA can be dangerous.
Why? Because the algorithm will naturally find it easier and cheaper to sell the £20 items. So, to hit its target, it will prioritise selling loads of your low-value products. You'll look at your report, see a great CPA and lots of conversions, and think you're smashing it. But when you check your bank account, your revenue is low because you've only sold the cheap stuff.
For most e-commerce businesses with a varied product catalogue, a Target ROAS (Return On Ad Spend) bid strategy is far superior. With tROAS, you tell Google, "For every £1 I spend on ads, I want to get at least £4 back in revenue." The system then stops caring about the *cost* of a conversion and starts caring about the *value* of a conversion. It will actively prioritise showing ads for your more expensive, higher-margin products because that's the easiest way for it to hit your ROAS target.
Switching from tCPA to tROAS can be a big change, and you need enough conversion data for it to work well, but it aligns the campaign directly with what actually matters: profitability, not just conversion volume. I remember one client who sold cleaning products with a range of prices; by focusing our strategy on return, we helped them achieve a 633% return and a 190% increase in revenue, precisely because the campaigns started optimising for value, not just the number of sales.
Before you make any big decisions, it's worth knowing your numbers inside and out. Specifically, what is a customer actually worth to you over their lifetime? This tells you how much you can really afford to spend to acquire them in the first place, both in peak and slow seasons.
Customer Lifetime Value (LTV) Calculator
This is the main advice I have for you:
To pull all this together, here’s a summary of the actionable steps you can take. This isn't just theory; it’s a framework we use to manage seasonal accounts every day.
| Action Item | Why It Matters | How to Implement |
|---|---|---|
| Stop worrying about the budget ceiling | The budget isn't the instruction; your tCPA/tROAS target is. An unspent budget during slow times is a sign the system is working correctly, not failing. | Set your budget to the maximum you're comfortable spending and then focus your attention on optimising your bid strategy and signals instead. |
| Proactively manage your tCPA target | This is how you guide the algorithm through seasonal shifts, preventing it from overspending or stopping delivery altogether. |
|
| Use Seasonality Adjustments | This gives the algorithm a direct heads-up about expected changes in conversion rate, making its bidding far more accurate. | In Google Ads, go to Tools & Settings > Bid Strategies > Advanced Controls. Create an adjustment for your slow period, estimating the expected drop in conversion rate. |
| Optimise Your Product Feed | For a feed-only campaign, this is your primary asset. A better feed means better targeting, better ads, and better results. It's non-negotiable. | Audit and improve product titles, descriptions, and images. Crucially, use Custom Labels to segment products by season, margin, or performance. |
| Consider switching to tROAS | If your product prices vary, tCPA can lead to optimising for low-value sales. tROAS aligns the campaign directly with revenue and profitability. | Ensure you have enough conversion data (50+ conversions in 30 days is a good rule of thumb). Then, create an experiment in Google Ads to test tROAS against your current tCPA setup. |
| Scale Budgets Slowly | Large, sudden budget increases will reset the learning phase, wrecking your performance for a week or more. Patience is everything here. | When you need to increase spend, do it gradually. A 15-20% increase every 3-4 days is a safe rule of thumb that avoids shocking the system. |
I know this is a lot to take in. Managing a PMax campaign for a seasonal business feels like it should be simple, but as you can see, there are quite a few moving parts. Getting it right can be the difference between a profitable year and a stressful one.
This kind of strategic oversight, digging into the data, and making proactive changes is where working with a specialist can really pay off. We're not just setting up campaigns; we're managing them as dynamic assets that need to adapt to the realities of your business cycle.
If you'd like to have a proper chat and a look through your account together, we offer a free, no-obligation strategy session. We can pinpoint exactly where the opportunities are and give you a clear plan of action. Let me know if that's something you'd be interested in.
Hope this helps clear things up a bit!
Regards,
Team @ Lukas Holschuh
