Google’s advice assumes you’re running big accounts. Upgrade to PMax. Switch to smart bidding. Migrate your DSA campaigns to AI Max. It all sounds great — until you realise it’s built for advertisers with large budgets, high conversion volumes, and plenty of search traffic to feed the machine.
Most of us aren’t managing those accounts.
I’ve been running Google Ads for 20 years — 14 of those as a freelancer — and the vast majority of accounts I work on are what I’d call low data accounts. They might be niche B2B businesses with limited monthly search volume (sustainability consultants, empty rates mitigation specialists, that kind of thing). They might be local businesses targeting a 10-mile radius. They might be accounts that simply don’t generate more than 30 conversions a month — that magic number Google says you need before switching to maximise conversions bidding. Or they might be brand new campaigns with zero history at all.
And here’s the thing: low data doesn’t have to mean poor performance. Our smallest active client spends £2 a day — 60 pounds a month — and generates more than 30 conversions. It’s not always a budget problem. It’s a data problem. And data problems require a different approach.
This is the framework I use.
The Low Data Optimisation Framework
There are four pillars:
- Preserve the signal — don’t dilute the data you do have
- Mine intent from your search terms — use what’s there to guide Google
- Protect the budget — stop paying for irrelevant clicks
- Prioritise big impact changes — small tweaks won’t move the needle
Let’s go through each one.
1. Preserve the Signal
With a high-data account, segmentation is your friend. More campaigns, tighter ad groups, better ad relevance, higher quality scores. Makes sense when you’ve got the volume to support it.
With a low-data account, that logic flips. Fewer campaigns means faster learning. Every time you fragment your account — splitting into more campaigns, more ad groups — you’re dividing an already small pool of conversion data into even smaller pots. Google can’t learn from data it doesn’t have enough of.
I see this a lot with new clients who’ve structured their campaigns to mirror their website. I worked with a garden supplies company that had a separate campaign for every product category they stocked: compost, tools, planters, seeds, you name it. They had a total daily budget of £10 across the whole account, which meant some campaigns were running on under a pound a day. They weren’t generating any meaningful data — and it turned out, people weren’t searching for specific product categories anyway. They were searching for “garden supplies near me” and “garden centre [location].”
We consolidated everything into a single, simpler structure. Conversion data started flowing, CPAs came down, client was happy.
The rule: resist the urge to segment until you have the data to justify it.
2. Mine Intent from Your Search Terms Report
If you’re in your search terms report at least once a month, you’re doing better than many. But if you’re only using it to find negatives, you’re leaving a lot of value on the table.
With low-data accounts, your search terms report is one of the few places where you can actually see intent signals — without needing conversion data to interpret them.
Here’s how to use it more actively:
Look for n-grams (repeating words). Spot a word appearing regularly across multiple search terms and ask what that tells you. “Free” appearing often? Negative it. “Supplier” or “services” appearing? That shows intent — lean into it. Can you build new keywords or ad copy around it?
Spot intent modifiers. Take a keyword like “HR consultant.” Typed on its own, there’s no real intent signal. But “HR consultant near me,” “HR consultant prices,” “HR consultant services” — those are different. Those people are looking to hire someone. Whereas “HR consultant definition” or “HR consultant salary”? Probably not your customer. Add them as negatives.
Find positives, not just negatives. The default mode in the search terms report is hunting for things to exclude. Flip that. Use it to find search terms that are performing well — even just getting clicks — and ask whether you can be more intentional about targeting them.
The rule: spend real time in your search terms report every week, and come out with more than just a list of negatives.
3. Protect the Budget
Low-data accounts often have low budgets too. Wasting spend on the wrong clicks isn’t just inefficient — it can genuinely derail performance by starving your good keywords of the data they need.
A few things that make a real difference:
Negative keywords from day one. Google’s recommendations won’t tell you to add negatives — there’s no incentive for them to do so. But adding a solid negative keyword list before you launch is one of the highest-value things you can do. Use AI to help build your initial list if you want, just sense-check it and be careful with match types. I’ve seen someone add “laptop” as a broad match negative on a laptop accessories account and wonder why impressions dropped overnight.
Avoid premature broad match. I’ve managed accounts for long enough to be suspicious of broad match by default — particularly for niche B2B where CPCs are already high. We had a client — a specialist HR consultancy — bidding on “employment law” as broad match. One of the resulting search terms was someone researching their own redundancy rights. That click cost £22. Nearly half their daily budget, gone on someone who was never going to become a customer. We went back to exact match.
That said: I do have accounts that perform well on broad match. The point isn’t never use it. The point is don’t rush into it before you understand how your keywords behave.
Don’t pay for the might-bes and maybes. A client once argued for bidding on competitor terms because their competitor was bidding on theirs. The problem? Their competitor was ten times the size with ten times the budget. Bidding on competitor terms means creating a new campaign, splitting your budget further, and often paying high CPCs for low-intent clicks. Be selective.
The rule: every pound of budget needs to be working. On a low-data account, wasted spend isn’t just wasteful — it actively hurts your learning.
4. Prioritise Big Impact Changes
On a big account, a 2% improvement in conversion rate compounds across thousands of conversions and makes a meaningful difference. On a low-data account, you won’t even see it. The numbers are too small.
That means your attention needs to go to the things that can move the needle significantly.
Landing pages. Is your page actually set up to convert? Social proof, clear headlines that address the customer’s pain point, a compelling offer, fast load time. This is CRO basics, but it matters more when you don’t have volume to compensate for a leaky funnel.
Intent alignment. Follow the thread from search term → ad copy → landing page. If those three things aren’t telling a consistent story, you’re losing people along the way.
Geography. We work with a lot of businesses in the North of England. When they try to target the whole UK, CPCs around London are high and conversion rates are low — they’re not local, so people don’t trust them. Excluding certain regions and focusing on areas where you actually win business can improve your conversion signal significantly.
Sequential testing. Parallel A/B testing doesn’t work well with low-data accounts — you rarely have enough volume to reach statistical significance in a useful timeframe. Instead, run version A for a set period (or until you hit a certain number of impressions or clicks), then switch to version B under the same conditions. Not perfect, but much better than running two versions simultaneously with 10 clicks between them.
Offline conversions. Just because an account is small doesn’t mean you shouldn’t be feeding lead quality data back to Google. If you don’t tell Google what a good conversion looks like, it’ll optimise for the cheapest and easiest ones — which are rarely your best customers. Yes, this is harder to implement when your client uses a notebook and their own memory instead of a CRM, but it’s worth pushing for.
The rule: ignore the incremental tweaks. Focus on the things that could actually shift results.
Putting It Together
Low data accounts can perform like big data accounts — but only if you manage them differently. The instinct to do more (more campaigns, more keywords, more segmentation) is usually wrong. The instinct to do less, but do it well, is usually right.
Preserve the signal you have. Mine your search terms for intent. Protect every pound of budget. And focus your energy on changes that matter.
One last thing: if you’re working with clients who have low-data accounts, manage expectations from the start. Tell them upfront that learning will take longer. Be honest about whether there’s enough search volume for the campaign to be viable. And learn to recognise when something isn’t going to work — and say so before you’ve spent their money finding out.
I presented this framework at HeroConf 2026 in Brighton. Further posts in this series dive deeper into each of the four pillars — search terms, budget protection, signal preservation, and prioritising big impact changes.
Leave a Reply