Why ROAS Can Improve While Your Business Gets Worse

ROAS is one of the easiest digital metrics to understand.

Spend £1. Make £4. ROAS = 4.

That simplicity is exactly why it can become dangerous.

A business can improve ROAS while revenue falls, new-customer acquisition weakens and the overall growth engine becomes less healthy.

Efficiency is not volume

Imagine a campaign spends £100,000 and generates £300,000 of attributed revenue. ROAS is 3.

The team tightens targets and removes weaker activity. Spend falls to £50,000 and attributed revenue falls to £200,000.

ROAS has improved to 4.

The campaign is more efficient. It is also generating £100,000 less reported revenue.

Whether that is good or bad depends on margin, incrementality, available demand and what the business needs.

The ratio cannot answer that by itself.

Platforms naturally find the easiest conversions

As targets become stricter, media can concentrate around audiences and searches most likely to convert.

That often includes existing demand: branded searches, returning visitors, people already familiar with the product or customers close to purchase.

The account looks cleaner. The uncomfortable question is how much new demand it is creating.

ROAS does not know your profit

Two products can have identical revenue and completely different contribution.

A campaign can also create revenue through discounting, expensive delivery or high-return products. The platform sees the conversion value you send it. The finance team sees what remains afterwards.

That gap matters.

ROAS is attribution, not causation

If a customer sees a social ad, searches the brand, clicks a paid result and purchases, more than one platform may claim the order.

I have worked in genuinely multichannel environments where print, email, paid media, affiliates and direct traffic all interact with the same customer. Platform data is useful for optimisation, but it is not an independent audit of who created the sale.

What I would look at alongside ROAS

I would add:

• total revenue and contribution
• spend and marginal return
• new-customer volume
• brand versus non-brand mix
• blended marketing efficiency
• conversion rate
• customer lifetime value where reliable
• incrementality testing where possible

The exact mix depends on the business.

ROAS is useful. Worshipping it is not.

I am not arguing for throwing ROAS away. It is a useful performance signal and an important control in many accounts.

I am arguing against allowing a channel efficiency ratio to become the definition of whether the business is growing well.

If ROAS is improving and everyone around the board table still feels worse, do not immediately assume the board is missing something.

The metric may only be telling part of the story.

If the platform reports look healthy but the commercial number does not, I can help you separate channel performance from the wider growth problem.