FEATURE3 October 2012

Channeling digital insight

Google knows huge amounts about what people do online. But for a full picture of the digital lives of consumers, it turns to its market insights team. Brian Tarran goes behind the scenes at the company’s UK headquarters.

For most people, Google serves as an unrivalled source of data about what people are doing online. If you’re a website publisher, Google Analytics will tell you most of what you need to know about who visits your site, how long they stay and what they do while they are there. If you don’t own a website or are just interested in finding out about the most popular web content of the day, Google’s Insights for Search allows you to see what the world is searching for, while Google Correlate allows you to match that search data with real-world events (sample fact: searches for ‘mittens’ go up during winter).

“We don’t really have a problem getting research listened to. Google understands the value of data, insight and evidence”

Sarah Everitt

Sarah Everitt

But if you happen to work for Google and you want to know more about people’s online behaviour, the market insight team is your best port of call. Market insights means understanding Google’s audiences and how they use Google’s properties and, more generally, what people do online, how they do it, how they interact with advertising and what effect that advertising has on their behaviour.

Sarah Everitt heads a team of five researchers in London, with others based in Hamburg, Paris, Madrid, Milan and Amsterdam. “We sit within the marketing organisation,” she says, “and our two key stakeholder groups are the sales and marketing teams.” It used to be an 80/20 split in favour of working with the sales team, Everitt says, but as the insights team has grown they are now being called upon to support the marketing team more often to help them understand “how consumers use Google’s products and how to position their marketing campaigns”.

Context is king

Plenty of people are surprised to discover that Google – with its vast stores of clickstream, search and analytics data – would actually have need of a research team, especially one that deals mainly in the types of respondent-level research projects often dismissed as unnecessary
in a world of Big Data.

“But we reach the metrics that the Google data source can’t reach,” explains Mark Riseley, group product marketing manager within the EMEA research team. “We, as a research team, often have to add more context than is available in the Google data to answer advertisers’ questions.”

One big question the team gets asked “all the time”, says research manager Jonny Protheroe, is how online and TV fit together – from both audience and advertising perspectives. One question in particular is proving very tricky to answer. That is, ‘To what extent does TV viewing drive online research and other online behaviour?’ Protheroe says: “In some ways, we can look at total Google search volumes to help answer that, but in other respects it’s useful to have a panel.”

“We’re probably under more pressure here than you would find in a lot of other companies to turn things around quickly”

Jonny Protheroe

Jonny Protheroe

The panel referred to is one currently being piloted with Kantar. The pilot involves developing an opt-in panel of UK households through which both TV and online media consumption will be measured using passive metering technology. Using the panel, Google should be able to get a better understanding of when and with what purpose people are using their mobiles, laptops and other devices while watching TV, or whether including YouTube in an advertising campaign brings in additional audience.

Protheroe is supported in his efforts here by Riseley, whose focus is on advertising effectiveness research. “With these panel initiatives, I’m really looking at developing benchmarks for different advertising formats,” he says. “Cross-media panels can tell you some interesting things about the reach overlap of different channels, but what marketers really need to help with their advertising planning is to have a sense of the relative effectiveness of, say, a TV ad or a YouTube pre-roll or an online banner ad.”

Google has a strong commercial interest in proving that online advertising works – and in developing new, even more effective ad formats. In April-June of this year, the company made $10.5bn in advertising revenue. For the last full year, 2011, that figure was $36.5bn. Of course, the more ad formats it has to offer customers, the more money it stands to make – and the growth potential is huge. Online advertising will only account for an estimated 17% of global ad spend this year, according to ZenithOptimedia forecasts. That’s $84.2bn of a total $485.9bn.

“You speak to the senior guys at YouTube and they honestly want to create ads that people love,” says Everitt. “The skippable ad format – which allows people to skip an ad after five seconds – is a really good example of that.” It’s built on the assumption that consumers will want to watch a fantastic piece of creative, but if they’ve seen it before or they don’t like it, they can go straight to their requested content after only a short delay.

It’s a nice idea, both in theory and execution – but nice ideas aren’t enough, Riseley says. “We need to prove it works in the most robust way possible. Search has always had a very high level of accountability and what we’re doing is trying to bring that type of accountability to other online ad formats.”

“Customer research journeys are getting shorter. They are becoming more direct. Consumers are becoming a little more savvy”

Louisa Middleton

Louisa Middleton


Google is both advertising network and advertiser. It has products to sell and it uses advertising to do that – and not just online advertising. “Google as an advertiser has the same needs as all the others,” says Richard Curling, an associate research manager at Google, who leads support for Google’s marketing teams. “Internally, we want to know how effectively and efficiently our advertising is working on different channels, whether it’s TV ads for our Chrome browser or Google Plus [a social network] or activity across the web…”

“And it’s about understanding the different audiences, too,” says Everitt. “Who’s using a product? What are they doing with it? How does it change or fit in with what they have done before?”

“A good example would be Chrome, and our job might be to understand whether the people we’re trying to talk to even know what a browser is so we can communicate with them in the right way,” says Riseley.

An advantage of Google’s dual role is that lessons learned from its own marketing campaigns or research projects can be channelled back to its advertising clients. Riseley says: “Effectively, we are dog-fooding the methods we are using with our advertisers. Often the marketing team will get the first go at some of the new research methods we’re trying out.”

Keeping up

The UK market insights team has been in place for about seven years now and in that time, Everitt feels she and her colleagues have become “respected partners” not only of the sales and marketing teams with which they work most closely, but with the product teams as well.

“We don’t really have a problem getting research listened to,” she says. “Google as a company understands the value of data, insight and evidence. The product teams will take on board the insights that our team delivers because they know we understand what consumers are doing, whether it’s because we’ve gone out there and run focus groups or we’ve taken data from third-party panels so we understand what people are doing throughout their online journey.”

If anything, Everitt says, the real problem is that “there are more people within Google that want research than we are necessarily able to deliver”.

“Internally, we want to know how effectively and efficiently our advertising is working on different channels”

Richard Curling

Richard Curling

“One of the biggest challenges for us is keeping up with the product teams,” says Protheroe. “We’re probably under more pressure here than you would find in a lot of other companies to turn things around quickly to make sure they are useful.”

Google is famous for its test-and-iterate product development philosophy: put it out there, see what works, then tweak it. “To some degree with all online technologies there is an element of measurement built in,” says Protheroe. “A simplistic example would be: if we can see that lots of people are using something, it’s probably quite good and quite useful. And if no one is, it probably isn’t.

“Those things speak for themselves but there is definitely a role for more traditional research techniques to come in there and add context to the hard numbers that the products themselves generate.”

In developing Gmail, for instance, Everitt says that alongside the usage stats, research was conducted to understand what would make someone switch email providers and what features would be most enticing. Five years on from its public release, Gmail now commands 31% of the webmail market – behind Yahoo Mail and Microsoft’s Hotmail, but it’s growing. Gmail’s unique visitors grew 17% year-on-year in June, according to ComScore, while Hotmail’s declined 4%.

Going mobile

The market insight team’s research priorities come from the very top of the organisation, says Protheroe, and this year there is a lot of interest in understanding how the mobile phone and other mobile devices are shaping the online experience. “There’s just a lack of information out there about how mobile affects behaviour: whether it is complementary or substitutive,” he says.

Key to understanding mobile is to understand where it fits into consumers’ daily lives and what its place is as the customer moves from researching a potential purchase to actually buying something. Overseeing that work is Louisa Middleton, a Google product marketing manager and clickstream research lead.

Customer research journeys are getting shorter, she says. “They are becoming more direct. Consumers are becoming a little more savvy.” This is a natural consequence of a maturing market, says Riseley. People are becoming more practised at finding what they want online.
But mobile devices are also helping reduce journey times. It’s not true for all categories, Everitt says, but one good example occurred in February this year. Google discovered that 62% of total US searches for chain restaurants that occurred on Valentine’s Day took place on high-end mobile devices or tablets. Mobile searches for flower-related terms also grew 227% from 7 to 14 February, while ‘click to call’ and ‘get directions’ links in ads proved very popular – up 560% and 514% respectively.

“Search has always had a high level of accountability We’re trying to bring that type of accountability to other online ad formats”

Mark Riseley

Mark Riseley

“Mobile really is a tool for research and purchase, especially when there is a sense of urgency,” says Everitt. Meanwhile, Curling points to surveys of smartphone owners that show the percentage of people reporting making a purchase through their device holding stable despite massive growth in the adoption of smartphones. “You would normally expect those sorts of behaviours to go down as a percentage of the total population as the population itself gets bigger,” he says.

“The fact that we’re seeing these things stay stable, or go up in some cases, is interesting,” he says. “It shows people are getting to grips with their mobile device and the functionality it has more quickly,” says Everitt.

But the growth of mobile brings with it its own challenges for advertising researchers, says Riseley. “One of the implications of mobile usage is that it makes it harder to look at the metrics produced by any one device and actually draw the right conclusions about whether something has worked or not,” he says. “You can have customer journeys that are starting on mobile but are finishing on a computer but if you are only looking at conversation rates from one device or the other you’re missing half the picture.”

User focus

The very existence of the market insights team is a testament to Google’s desire to see the full picture of online consumer behaviour. Despite the petabytes of behavioural data stored on its servers, Google still recognises that it needs to know more about why people do what they do and set that in the context of their offline lives.

It’s tempting to think that having grown up with the internet, and becoming one of the world’s biggest businesses in the process, Google knows instinctively what works and what doesn’t for an online audience. But like all other companies, Google also has to feel its way to success, innovating and adapting as it goes, steered by the evidence in front of it.

“Focus on the user and all else will follow” is the first and perhaps most important of Google’s ten principles – and the market insights team truly embodies it.

TrueView: ad engagement

YouTube’s skippable ad format, TrueView, launched in December 2010, allowing YouTube users to skip pre-roll ads after the first five seconds. “It’s a bit counter-intuitive,” wrote Techcrunch’s Jason Kincaid at the time. “Advertisers are constantly looking for ways to force you to watch an ad, after all.”

Clearly Google had a job to do to convince brand owners of the benefit of this format, so the UK market insight team commissioned research from Ipsos MediaCT and Innerscope to determine the impact of TrueView on user engagement and brand uplift.

In a lab, biometric measurements were used to record physiological responses to three different ad formats: TrueView, a non-skippable pre-roll and no pre-roll. The research found that on average people who watched TrueView ads showed 75% greater levels of engagement with the ad than those who saw the same ad as a standard pre-roll.

In addition, Google said users who watched a TrueView ad were 40% more engaged during the first five seconds than users who saw a standard pre-roll, while even those choosing not to watch the ads were 30% more engaged during the first five seconds.

So skipped ads could still deliver advertising value – and best of all, those five seconds are free. Advertisers are only charged when an ad is watched in full or for more than 30 seconds, whichever occurs first.