FEATURE1 July 2009

Getting the measure of ROI on ad spend


How do you go about providing a successful solution for measuring ad ROI? Robert Bain finds that staying focused on your own ROI is a good start.


?Everything seems to be conspiring to make advertisers more anxious to know what they’re getting back for what they spend. Media are becoming more fragmented, consumers more elusive, the economic outlook is relentlessly bleak and the inherent measurability of the internet is making other media look like a risky bet.

A recent study by Lenskold Group – a provider of marketing ROI tools – said 65% of marketers surveyed were being asked to show potential ROI before being assigned a budget, and 79% said the need to measure effectiveness was greater this year than last, because of deteriorating economic conditions.

As a result ROI has become a buzzword among researchers, both in the sense of ‘How can we measure it for our clients?’ and ‘How can we provide it as researchers?’

In Advertising Works, which brings together the winning case studies from the 2008 IPA Effectiveness Awards, Les Binet of DDB Matrix offers some guidance on how to calculate payback. Binet writes that, since the awards exist to prove advertising works, the study authors seem “remarkably ill-informed about how to measure payback and return on investment”. He criticises those who calculate ROI incorrectly, those who neglect to measure it at all and those who misuse the term to refer to other measures of things like awareness and direct response rates.

Binet’s assessment suggests that – to put it mildly – there is plenty of room for improvement on the part of media researchers.

In 2004 media measurement giants Nielsen and Arbitron began Project Apollo, a major trial aimed at taking the guesswork out of measuring advertising ROI through a single panel capturing participants’ viewing and purchase behaviour. The findings were lauded as groundbreaking. Unfortunately the sheer cost made the project unsustainable and it was abandoned last year.

“TV ads have been bought and sold for years on the basis of sex-age demographics, which were never anything more than a proxy for purchase behaviour”

But a consensus remained that single source was something that ‘had to happen’. After all, TV ads have been bought and sold for years on the basis of sex-age demographics, which were never anything more than a proxy for purchase behaviour. The potential to work with actual purchase behaviour is irresistible.

At the same time that the industry was waving goodbye to Apollo, a startup called TRA unveiled its own ad ROI system, which sidestepped the challenge of setting up a panel of its own by making use of information that was already out there: set-top box data from more than a million TiVo and cable subscribers, and purchase data from the frequent-shopper card programmes of six retailers with a total of 1,000 outlets. This gives them a combined sample of 370,000 homes for whom they have access to viewing and shopping data – far more than Apollo’s 5,000.

Arbitron’s then CEO Steve Morris said TRA’s alternative lacked the “granularity and person-level data” of Apollo, but his criticisms sound rather hollow given that Apollo was a commercial failure and TRA is now busy signing up clients and preparing to expand its offering later this year. In fact Arbitron itself has now come on board as one of TRA’s backers, leading a $13.5m investment round in May. Clients include broadcasters, CPG manufacturers and media agencies.

Thirst for a solution
Research asked TRA’s co-founder and CEO Mark Lieberman what allowed his company to succeed where the big players failed.

“Project Apollo blazed a great trail for us and became a bellwether, if you will, for market demand,” said Lieberman. “Advertisers were thirsty for a single source solution.”

He believes it was a combination of technology and the growing pressure on TV ad revenue from the digital world that made it possible for TRA to grasp the opportunity.

“There’s a loss of confidence in TV compared to other digital media because the measurement systems out there have not kept pace with the digital revolution in TV”

Mark Lieberman, TRA

“Ten years ago this would not have been possible. The number one thing that enabled us to do this was the software that needs to be downloaded into the set-top boxes across the country, so that we have a significant sample size on the TV side. Number two is that there are huge data sets out there in the form of frequent shopper card databases, which is where we get our purchase data. The last thing that enables us to be successful is really the marketplace. What’s happened in the last three or four years is an acceleration of ad dollars moving out of TV and into the internet. TV has been flat and is projected to be flat for the next couple of years, whereas the internet has been seeing double-digit growth. There’s a loss of confidence in TV compared to other digital media because the measurement systems out there have not kept pace with the digital revolution in TV – the current measurement system is still analogue-based.”

A question of direction
Lieberman said Apollo was heading in the right direction, but going about it the wrong way. “Firstly, Project Apollo was a joint venture, and joint ventures are difficult,” he said. “Nielsen and Arbitron are arguably competitors, certainly now more so than ever, and I think that always makes it difficult, running a JV between two companies that compete. We are a company backed by investors, they are looking for ROI, so we are very focused on making the best product we can to become profitable.

“Secondly, in terms of execution, Apollo had a tiny sample size of 5,000. Our sample size is 370,000 single-source households – that’s 70 times the size of Project Apollo. If you’re a brand, having a large sample becomes very important, especially if you want to analyse by geography, and 5,000 just didn’t cut it. The way I look at it, they were a mile wide and an inch deep. We’re probably a foot wide and a mile deep.

“Thirdly the prices they wanted to charge to justify the huge costs because of their inefficient solution just was not a cost that the marketplace was willing to bear. We have a much more efficient business model that allows us to charge a lot less. I believe that being an entrepreneurial young start-up, laser-focused on leading with what we know we can succeed with, gives us the opportunity to achieve Apollo’s vision in the long term.”

As it turns out, then, TRA’s edge came down to a focus on its own ROI. The company’s name stands for True ROI Accountability and, Lieberman enjoys telling clients, “ROI is our middle name.” And with Arbitron now counted among its investors, the firm hopes to benefit from the more detailed learnings from Project Apollo. In Q3 this year the firm plans to roll out into a major additional purchase category and in Q4 it’s going to introduce more media data. It hopes to turn its first profit next year.

Single-source data isn’t going to solve all of an advertiser’s problems by itself. Les Binet’s advice on measuring ROI comes with a further warning on how that data is used. He reminds marketers not to confuse a sales increase with incremental sales, or to look at revenue rather than profit as a measure of ROI. He also encourages them to keep an eye on the longer and broader effects of marketing in supporting higher prices, reducing business costs, creating new opportunities and changing market expectations.

Finding ways to answers these questions could open up fertile ground for researchers looking to provide clarity in a time of great uncertainty. When asked if he expects the single-source approach to spread into other areas of research, Lieberman responds: “You mean will other companies copy what we’re doing? I suspect there will be others that will do similar things. It’s the most sincere form of flattery, right?”

1 Comment

11 years ago

plz tell how to calculate the print media ad can be measured & how can one know the effectiveness of the ad

Like Report