This is part of a series on the Low Data Google Ads Optimisation Framework. Start with the flagship post for the full picture.
There’s a certain type of Google Ads optimisation that feels productive but achieves almost nothing on a low data account. Tweaking a bid adjustment by 5%. Shuffling ad copy variants that have four impressions between them. A/B testing a headline change when you’re getting 30 clicks a month.
On a big account, marginal gains stack up. A 2% improvement in conversion rate across thousands of conversions is genuinely meaningful. On a low data account, that same 2% improvement is invisible — too small to show up in the data, too small to feel in the results, and too easy to attribute to normal statistical noise rather than anything you actually did.
This isn’t an argument for doing nothing. It’s an argument for doing the right things — the ones that can shift performance significantly, not incrementally.
Start with your landing page
If there’s one place to look first for meaningful improvement on a low data account, it’s the landing page. No amount of clever keyword strategy compensates for sending hard-won clicks to a page that doesn’t convert.
Ask yourself honestly:
Does the page make a compelling case? Is there social proof — testimonials, case studies, accreditations, client logos? Do the headlines speak directly to the problem the customer is trying to solve, or do they just describe what you do? Have you clearly articulated the pain point before you present the solution?
Is the offer attractive and clear? Someone landing on your page should immediately understand what you’re offering, why it’s worth their time, and what they need to do next. If any of those three things require effort to figure out, you’re losing people.
Does the intent carry through? This one is easy to overlook. The thread from search term to ad copy to landing page needs to be consistent. If someone searches “emergency garage door repair,” clicks an ad that promises fast local repairs, and lands on a generic homepage about home improvements — they’ll leave. The intent they arrived with needs to be met immediately.
This is CRO fundamentals, but it has an outsized effect on low data accounts precisely because every click is precious. A landing page that converts at 8% instead of 4% doubles your results without changing a single keyword.
Look at your geography
Geography is an underused lever, and it’s one that can make a significant difference — particularly for businesses that serve a national or regional market but convert better in certain areas.
I work with a lot of businesses based in the North of England. Many of them technically serve the whole of the UK, and they want to target the whole of the UK. The problem is that when they do, CPCs in and around London tend to be considerably higher — more competition, higher cost per click — and conversion rates tend to be lower, because customers in London aren’t necessarily drawn to a business based in Leeds or Manchester that they’ve never heard of. Local trust matters.
By reviewing performance by location and excluding areas where CPCs are high and conversion rates are low, we can redirect that budget towards areas where the business actually wins. It’s a simple change with a real impact on the efficiency of the account — and it improves the quality of the conversion signal going to Google at the same time.
It’s worth doing this analysis even if the results surprise you. Sometimes the geographic performance data tells you something useful about where your business actually resonates.
Sequential testing: the realistic alternative to A/B testing
Parallel A/B testing is the gold standard. Run version A and version B simultaneously, in identical conditions, and let the data tell you which performs better. Clean, scientific, reliable.
It also doesn’t work well for most low data accounts.
The problem is statistical significance. To get a reliable result from a split test, you need enough conversions on each variant for the difference to mean something. On an account generating 20 conversions a month, split between two landing page variants, you might wait three to six months for a clear answer — by which point the business context has changed, seasonality has come and gone, and you’re not sure the comparison is valid any more.
The practical alternative is sequential testing. Run version A for a set period — or better, until you reach a defined number of impressions or clicks — then switch to version B and run it under the same conditions. Compare the results when both have accumulated the same exposure.
It’s not perfect. Seasonal variation between the two periods can affect the comparison. External factors might influence one period more than another. But it’s far better than running a split test with insufficient volume and drawing conclusions from noise, or worse, never testing at all because the conditions for proper testing never materialise.
Pick an important page, define your metric, set your threshold (500 clicks, say, or 1,000 impressions), and start. Sequential testing with a defined methodology beats the endless intention to A/B test properly one day.
Offline conversions: close the loop
This one gets overlooked more than it should, especially on smaller accounts.
Google’s automated bidding optimises towards the conversions you tell it about. If you’re only tracking form fills or phone calls, that’s what it optimises for — regardless of whether those leads are actually any good. If 40% of your form fills are tyre-kickers and only 60% become real customers, Google doesn’t know that. It’s optimising towards a metric that includes a lot of waste.
Feeding offline conversion data back into Google — which leads actually (or could potentially) turn into paying customers, and ideally what they are worth — gives Google a much more accurate target to optimise towards. It’s the difference between telling Google “get me leads” and “get me customers.”
I’ll be honest: this is harder to implement with smaller clients. In my experience, only a handful of the clients I work with have a proper CRM system. Some use spreadsheets. Some use notebooks. One client, memorably, stores all their customer data in their own head and shares none of it.
But it’s worth pushing for, even in basic form. A simple spreadsheet of converted leads uploaded monthly is better than nothing. If Google is going to optimise your account, give it the best information you can about what a good outcome actually looks like.
The things not worth your time on a low data account
Just as important as knowing where to focus is knowing what to deprioritise.
Minor bid adjustments. Small percentage tweaks to bids on low-volume campaigns won’t produce a measurable effect. If you’re going to adjust bids, make changes that are significant enough to actually change behaviour.
Ad copy testing with insufficient volume. If an ad variant has 15 impressions, you have no idea whether it’s good or bad. Wait until you have meaningful data before drawing conclusions — or at least treat early signals as directional rather than definitive.
Over-optimising quality scores on low-traffic keywords. Quality score matters, but chasing it on keywords that get five searches a month is not the best use of your time on a low data account.
Endless restructuring. As covered in the Preserve the Signal post, restructuring a low data account repeatedly resets the learning phase and dilutes your data further. If the structure isn’t the problem, don’t change it.
Focus where it counts
The discipline this pillar requires is the discipline of prioritisation — being honest about which tasks will actually move the needle and which ones just feel like productive optimisation.
On a low data account, the big wins tend to come from outside the platform as much as within it. Your landing page, your offer, your geographic focus, the quality of data you’re feeding back to Google. These aren’t as immediately satisfying as going into the account and making changes, but they’re where the real leverage is.
Make the big changes. Let the small ones wait until you have the volume to justify them.
Better decisions with less data start with knowing which decisions actually matter.
This is the final post in the Low Data Google Ads Optimisation Framework series. If you found it useful, the other posts cover Preserve the Signal, Mine Intent from Your Search Terms, and Protect the Budget. Or start with the flagship post for the full framework overview.
Leave a Reply