Case study · Saatchi & Saatchi · 2016

The client was cutting print because it proved nothing

Vehicle rental · shifting a print budget to measured search · Indian Ocean · 2016

Illustration: a folded newspaper in grey on the left, a search results screen in orange on the right, a budget arrow moving from one to the other

The client was not looking for a digital agency. He explained that he was cutting his print spend because it delivered very few results. That is the most useful sentence an advertiser can say, provided you know what to do with it.

We did not answer with a recommendation. We answered with a short paid search test campaign built around one question: does this budget produce something measurable, yes or no.

The best ad group came in at 6.73% click-through at an average cost per click of 0.61 euro, and a bounce rate of only 45%, meaning more than one visitor in two carried on. Three metrics, one message: this budget exists and it works.

01 · Context

A vehicle rental operator with two brands in the same island market, a near non-existent social presence, and most of its budget in local print. The sector is seasonal, the purchase decision happens in minutes on a search engine, and competition is regional. The digital starting point was close to zero on both brands, which is good news: everything measured afterwards is attributable.

The constraint

Taking budget from an incumbent medium is not a matter of conviction, it is a matter of proof, and proof costs money you do not have yet. So you need a test budget small enough to be granted without heavy arbitration, and large enough to produce a readable result. Add seasonality, which makes any period-to-period comparison arguable, and the fact that a client already disappointed by one medium is wary of the next.

03 · The system

1 · Start from the client's sentence, not yours.

He said print gave few results. So the way in was not to sell digital, it was to offer a measuring instrument. The difference is that the second proposal is hard to turn down.

2 · Ask for a test, not an annual budget.

A short campaign, a narrow scope, a decision at the end. A test that fails costs little and earns credibility. An annual plan that fails costs the account.

3 · Choose ad groups that answer the question.

The best balance of volume and quality came from an ad group carrying a specific offer with its discount, not from a brand query. A brand query collects traffic that would have come anyway. The test only proves something if it is about incremental demand.

4 · Look at the bounce as much as the click.

A high click-through rate on a poorly targeted ad produces traffic that leaves. The bounce at 45% is what turns the result into an argument: people did not only come, they stayed.

5 · Lay out a complete architecture, not just a campaign.

The pre-sale covered strategy, paid and organic search, social, video, events and a distribution partnership. The test campaign was not the offer, it was the way into a broader offer that became credible once the proof was made.

04 · Results

2016 test campaign, pitch stage.

6.73% click-through on the best ad group
Average cost per click of 0.61 euro across the test campaign
Bounce rate of 45% on the strongest ad group, so more than one visitor in two carried on
Best volume and quality pairing achieved on a specific offer, not on a brand query
Full architecture set out at pitch: strategy, paid and organic search, social, video, events, distribution partnership
Digital starting point close to zero on both brands, so the effect is fully attributable

05 · What it proves

An advertiser telling you their medium proves nothing is handing you the argument. The right answer is not to sell another medium, it is to sell the proof. A well chosen test campaign costs less than a thirty-page recommendation and settles the question in three weeks. If you cannot prove a pound is working, somebody will eventually give it back to print.