Rebuilding a Campaign Around Less Data

What the old approach assumed

The habits most advertisers learned were built on abundant signal.

Narrow audiences, many small campaigns, tight exclusions, and heavy retargeting all rely on being able to identify individuals reliably across properties and to observe what they did afterwards.

As that identification becomes partial, each of those techniques degrades, and some of them degrade into being actively counterproductive.

The instinct when results soften is to target more precisely. That is the wrong direction now, and it is worth understanding why before rebuilding anything.

Why narrow now works worse

A campaign learns from conversions. Below a certain number in a given period, the system has too little to work with and its optimisation is close to guessing.

Splitting a budget across eight small campaigns divides the conversions eight ways, so none of them accumulates enough to learn from.

That was survivable when signal was plentiful. With fewer observable conversions to begin with, dividing the remainder is how an account stops improving altogether.

Consolidation is therefore not merely tidier. It is what allows anything to be learned at all.

Consolidate first

• Fewer campaigns, each with a meaningful budget.
• Fewer audience segments within each.
• One clear objective per campaign.
• Enough weekly conversions for the system to optimise.
• Fewer exclusions, which fragment audiences further.

As a rough guide, a campaign producing fewer than about a dozen conversions a week is unlikely to be optimising meaningfully, and merging it into another is usually better than tuning it.

Broader audiences, not vaguer ones

A distinction worth holding, because broad is easily misread as careless.

Broad means giving the system a large pool and letting it find the responsive people within it, using the signals it still has.

Vague means advertising to everybody with no relevance, which remains a waste.

In practice, a wide geographic area with a sensible age range and no interest targeting frequently outperforms a tightly defined audience assembled from a dozen interests, because the tight definition was based on inference that is now less reliable.

Let the constraint be your service area and your offer rather than a list of attributes.

The bits that still work

Not everything degraded, and knowing what held up prevents overcorrection.

Search advertising, where somebody types what they want and clicks, is largely intact. Intent is stated rather than inferred, and the conversion happens on your own site.

Geographic targeting works, since it does not depend on following anybody.

Your own customer lists, uploaded to a platform, still function, and they are yours rather than borrowed.

And on-site measurement of what happens after somebody arrives is unaffected, which means landing page performance can still be assessed properly.

Move effort to the offer

The strategic consequence, and the one most likely to be resisted.

When targeting was precise, it was reasonable to spend most of your attention there. As precision falls, the leverage moves to what everybody who sees the ad encounters: the offer, the message, and the page they land on.

A better landing page improves results for every visitor regardless of how they were found, and no platform change can take that away.

Businesses that had been optimising audience settings weekly and had not rewritten their landing page in two years now have that priority the wrong way round.

A worked example

A home improvement company ran eleven campaigns split by service and by district, each with a small daily budget.

Total conversions were adequate, spread across eleven campaigns they were not, and nine of the eleven were producing fewer than five a week.

They consolidated to three: one per major service, each covering the whole service area, with interest targeting removed entirely.

They then spent the time freed up rewriting the two landing pages those campaigns pointed at, adding prices, photographs of their own completed work, and an honest section about what the job involves.

Cost per enquiry fell by roughly a quarter over two months.

They could no longer say which district an enquiry came from within the platform, which they had believed mattered and which turned out to be answerable from their own records anyway.

Give it time to settle

A practical point that catches people mid-rebuild.

Consolidating resets the learning on a campaign, so the first week or two after a change is not representative and will often look worse.

Judging it after five days and reverting is how accounts end up in permanent churn, never stable long enough to perform.

Make the change, leave it alone for at least two weeks, and compare a full month against the month before.

Note the date of every structural change somewhere you will find it later, because in six weeks nobody remembers what was altered when.

Measure what you can still own

The complement to all of this, and it is not a platform setting.

Ask every enquiry how they heard of you and record the answer somewhere countable.

Use a distinct phone number or landing address for a campaign where knowing matters enough to justify it.

Compare total monthly spend against total monthly enquiries, which is unaffected by any of this and answers the question you actually have.

These are cruder than what platforms offered three years ago and considerably more durable, since nothing outside your business can withdraw them.

The counter-case

Broad and consolidated is not universally right, and applying it mechanically causes its own problems.

Where services have genuinely different economics, a plumbing emergency against a bathroom refit, merging them means optimising toward whichever converts more easily rather than whichever is worth more.

Where budgets are large enough that even split campaigns clear the learning threshold comfortably, the argument for consolidation weakens considerably.

And a business with a genuinely narrow audience, selling to one profession in one region, may have no broader pool available to widen into.

The principle is to give each campaign enough conversions to learn from. Consolidation is the usual way to achieve that, not the goal itself.

The rebuild

1. Count weekly conversions per campaign.
2. Merge anything below about a dozen.
3. Remove interest targeting and widen the audience.
4. Keep geography and your own customer lists.
5. Rewrite the landing pages with the time freed up.
6. Leave it two weeks before judging.
7. Record the date of every structural change.

Step five is where the return actually is, and it is the step that gets postponed.

Where the money is really decided is covered in the landing page decides the cost.

Frequently asked questions

West Coast Media Solutions Inc. provides web design, web development, hosting, digital marketing, and business consulting to organisations across Canada, drawing on more than twenty-five years in the field.