The best digital campaign I worked on was in 2009.

I found an old campaign report the other day. It’s from 2009, when I worked in-house at an Australian internet provider, and it covers the launch of a new product: an all-in-one box for home internet and phone.

I expected to wince. Seventeen years is a long time in digital, and we’ve had every one of those years to get better at this.

I didn’t wince. I ended up wondering what happened.

The ads had jokes in them

The product was a black box that sat next to your phone socket. There’s only so much you can say about that, so the ads didn’t try to say it straight.

The product had a short, friendly first name, the kind a person has. One set of banners was a fill-in-the-blanks: empty spaces where the name went, followed by the surname of a famous person who shared it. You did the rest yourself.

Another said it was so versatile it could tell you jokes, and then told you one. Another listed the colours it came in: Jet Black, Emo Black, Jack Black.

_ _ _ Famous Surname You fill in the rest. Wifi. Phone. Voicemail. Nose trimmer.* * Joking Available in: Jet Black Emo Black Jack Black
Redrawn from memory, with the product’s name left out. A banner with a joke in it cost no more to run than a dull one.

When the second model came along, the concepts kept going. “The Second Coming.” A straight-faced feature list that read “Wifi. Phone. Voicemail. Nose trimmer.” with an asterisk, and “Joking” in the small print. And my favourite line in the whole deck: “NBN ready (for, um, when the NBN is ready).”

None of this was complicated. It was a product, a personality and somebody who could write.

The billboards knew where they were

The creative agency did something I’ve rarely seen since. When big billboard sites were being bought, they went and spent time in those places. They talked to locals. Then they wrote the billboard for that street, with things in it that only people who lived or drove there would get.

That’s what relevance looked like before it meant a targeting setting. Someone had to go and stand there.

A line only this street would understand. Written after an afternoon spent standing here.
A sketch of the idea, not one of the real billboards.

We put the ads in odd places, on purpose

The media plan wasn’t timid either. Alongside search and the news and technology sites, the campaign ran inside video games, in the Facebook news feed as video, on YouTube’s search results, and as a takeover that came out over the top of a news homepage.

Some of it was there to sell. Most of it was there to be noticed by the right people in a place they didn’t expect an internet provider to turn up.

And then we measured the people who didn’t click

This is the part that stopped me.

About three in every thousand ads were clicked. That’s normal, then and now. A report that only counted clicks would have had nothing to say about the other 997.

So the campaign was also measured with a control and exposed study. Over the weeks it ran, more than a thousand people on the sites carrying the ads were surveyed. Some had been shown the campaign and some hadn’t. Otherwise they were the same sort of people, asked the same questions at the same time. Any difference between the two groups was the advertising.

The difference was large.

  • Heard of the product: among people who hadn’t seen the ads, somewhere between one in eight and one in five. Among people who had, about 12 or 13 points higher.
  • Thought well of it: up about 7 points.
  • Would consider buying one: up about 6 or 7 points.
  • How many times they needed to see it: once was enough to show up in the numbers. More exposures kept building on it, and there was no sign of people getting sick of the ads even at eight or more.
Had heard of the product about 15% about 28% Thought well of it about 7% about 14% Would consider buying it about 10% about 17%
Approximate figures from the study. Outlined bars are people who weren’t shown the ads. Blue bars are people who were.

It also told us where the campaign worked hardest, which ad shapes pulled their weight and which didn’t, and which competitors’ customers were most open to it. That’s a planning tool for the next campaign, and none of it came from a click.

In the years of agency work since, I can count on very few fingers the times I’ve seen a study like that commissioned.

So what happened?

I don’t think the people got worse. I think the incentives changed.

Digital became the place where you could trace a sale back to a click, and everything bent towards that. Agencies who were brilliant at traditional advertising, and who had just proved they could bring it online, came to believe the brand work belonged on television and billboards, and that digital was for the retail message: product, price, button.

At the same time advertisers fell for targeting. The thinking goes that if you can aim at people to within an inch of their lives, the ad itself hardly matters. So the effort goes into who sees it, and almost none into what they see.

The result is a sameness you can check for yourself. Scroll any feed and count the ads that are a picture of the product, a line about the product and a button. They’re accurately delivered and they make no impression.

We got very good at tracking an impression and forgot about making one.

The other 99 per cent, again

I’ve written before about how most of what you buy in digital is an impression nobody clicks, and how little of the reporting is about those impressions. This campaign is the same point from the other side. In 2009, with far cruder tools, we put real effort into what the 99 per cent saw, and then went and found out what it did to them.

To be fair, great work still gets made. I see it, and I’m glad every time. It’s just rarer than it should be in the channels where most of the money now goes.

The more spirit you put into your digital advertising, the more it stands out, and that includes the ads that are there to sell. It’s also easier to stand out now than it was then, because so much around you looks the same.

What I’d take from 2009

  1. Give the product a personality, and let someone who can write loose on it. A fill-in-the-blanks banner cost no more to run than a dull one.
  2. Make relevance by hand sometimes. Go and stand where the ad will be, or sit in the feed it will appear in, and write for that.
  3. Spend as much on the message as on the targeting. Reaching exactly the right person with nothing to say is still nothing.
  4. Measure the people who saw it and didn’t click. A control group and a survey did it in 2009. There are more ways now, not fewer.

I’d like to think the next seventeen years go better than the last. A good start would be getting back to where we were.

Sound familiar?

Tell me what’s going on. I’ll reply with questions or a scoped proposal.