Bad data is costing businesses at least 30% of revenues, according to research by analysts Ovum in 2012, while in December 2013 a Demand Gen Report revealed that more than 62% of organisations rely on marketing or prospect data that is 20 to 40% incomplete or inaccurate. But as data continues to become more central – and plentiful – within brands’ media buying, is it becoming more reliable?
“Marketing organisations are getting much more data-driven, and as a pure play we are literally swimming in data, so we have to make sure that the data we are using enables us to make the right decisions,” says Ben Carter, marketing director at online marketplace Notonthehighstreet.com.
He is not alone in continually battling to harness the wealth of media data available to brands today, identifying what is useful and what is noise. Claire Round, director of marketing and brand at English National Opera (ENO), says: “I made the mistake of trying to collect every piece of data I could, but that makes it impossible to make a decision, so you have to work out which data you actually need. It’s not just about data for data’s sake.”
Multiple data sources
Part of the answer lies in ensuring that multiple data sources do not exist in separate silos. Online clothing retailer Surfdome, which works with Quantcast, has recently invested in this area, appointing a data specialist earlier this year. Head of paid channels Matt Slingsby says: “[The data specialist] has put in place a big data analytics package that consolidates all our data sources into one, giving us what is approaching one single version of the truth, so everyone is working on the same page and the same data, meaning insights become trustworthy and easily actionable.” The retailer uses real-time sales data, lifetime data and more recently gender data.
One finding from its paid media activity was that the company was spending more money advertising to women than men. “We realised men provided more value to us as a company because return rates were lower and conversion rates were higher. We started re-evaluating where we put that money based on lifetime value, but from a gender perspective,” explains Slingsby.
He adds that although it is always important to query any data, the company is “at a stage where all the influence-makers are happy with the data and insights, and we have enough trust to be able to say ‘ok, let’s put £400k into this [media buy]’.”
Publisher Northern & Shell, whose titles include Express newspapers and OK! magazine, has also brought together disparate data sources, partnering with technology vendor Amido to create a single view that improves the reliability of its audience data as a media owner. The days when a print publisher could rely solely on its ABC circulation figure to sell ads seem to be numbered.
“The only way we can actually make a difference from a data point of view is to put people into audience segments and we can then serve them personalised ads on the fly when they are on the website”
Deni Boncheva, Northern & Shell
“I inherited a lot of digital platforms that didn’t really talk to each other,” says head of CRM Deni Boncheva. “The data didn’t work together. So we have been addressing how to get people to register, and how to combine all the data across our different brands into one data pot and make sense of it – online and offline.”
ENO has also been able to aggregate its online and offline data, working with Total Media and using the QlikView data visualisation platform. ENO’s Round says: “[QlikView] takes data feeds of all our information including historic sales information, social media information, email engagement and comparative Nielsen Media Research data, enabling us to start visualising the sales pattern and how it benchmarks against other things in the marketplace.”
The company has recently moved more of its media budget into digital, which is now around 50%. “We put tracking tags all the way through the purchasing journey, and have increased that spend because we have seen a really good conversion in sales,” explains Round. Not that tracking tags are always reliable. “If something happens, such as a software upgrade, the tracking tags don’t stick.” It is one of a number of causes of potential discrepancies. “We look at historical analysis too, but you have to ensure sometimes that you are comparing like for like.”
However, ENO does learn lessons from each campaign. “We are just reviewing our recent La Bohème campaign because we have La Traviata coming up in February and they have fairly similar audience profiles,” says Round. “We will look at what we can learn from that campaign in terms of optimising the next one from the start.”
As a London venue, ENO’s offline media spend includes out-of-home (OOH) advertising on the London Underground, as well as radio. Round says the organisation segments its customer data, as well as using information from Kantar Media TGI, to give it a sense of its audiences’ cultural engagement and behaviour. “For example, if we are looking at one target audience and a significant percentage are within the Classic FM [audience demographic], that seems like a sensible place to put that campaign.”
Notonthehighstreet also uses offline advertising. In September it ran its first OOH campaign, promoting its offering around London tube stations. It worked with media agency M2M to identify the audience it was targeting, as well as using ‘mosaic’ data from TGI to profile its customer base and combine this with attitudinal segmentation. “We pulled all of those together and identified the right media to reach those people,” says Carter. “For example, we knew we didn’t want to be in Oxford Circus because it is a big tourist station and there would be a lot of wastage – we wanted to go after commuter stations where we knew people would give more consideration to the message.” The campaign prompted an uplift in demand in what is usually a quiet trading period.
The retailer will also increase its focus on TV advertising next year. It has used econometrics to show that TV has been working, but this has also been supported by online data. “We can also see the TV success through the correlation to search impressions as well as overall website traffic when our ads are on TV,” says Carter. But the company plans to take this further: “We want to drill down to what is the most effective and efficient media buy, so we will do spot by spot analysis. We have a bespoke audience who we go after, which is a combination of our target audience and some attitudinal segments that we have identified, and that is cross checked against TGI data to say what their media consumption behaviour will be like. For example, we know our audience tends to start watching TV later in the evening.”
Change in emphasis
One key shift online is the move from reliance on third-party data to an increased emphasis on gathering first-party data, whether from the brand itself or from a publisher’s direct relationship with its audiences. Northern & Shell can track every customer touchpoint on its website, from browsing behaviour to registration, login and competitions, and this is combined with social media data, allowing it to segment customers. “For example, if someone is on the Daily Express website and we also know from their Facebook interests that they like a certain political party and regularly comment on articles about politics, we put them into a segment called ‘politics’,” says Boncheva.
“Programmatic advertising already accounts for more than half the revenues for many online publishers and we expect this to grow further”
Rakesh Patel, Auto Trader
The data is also real time, enabling the publisher to serve programmatic advertising. “The only way we can actually make a difference from a data point of view is to put people into audience segments and we can then serve them personalised ads on the fly when they are on the website,” she explains.
Automotive website Auto Trader, which is involved in the media buying market both as a publisher and advertiser (working with Vizeum), is also seeing a rise in programmatic advertising. Digital sales director Rakesh Patel says: “Programmatic advertising already accounts for more than half the revenues for many online publishers and we expect this to grow further.” He adds that Auto Trader offers its advertisers, which include Hyundai, Kia and Mercedes-Benz, detailed consumer segments based on first-party data to be targeted in real time. As an advertiser itself, 70% of Auto Trader’s digital investments are in programmatic advertising.
This trend towards programmatic is further undermining the importance of third-party data online. Surfdome’s Slingsby says it now uses largely proprietary data. “Everything we work on is raw data and we pump it into one system.” The retailer used to rely heavily on Google Analytics but Slingsby says it has moved away from this as he is unsure of its accuracy. “For trends such as day-to-day comparisons and year-on-year, it is absolutely fine, but in terms of pure decision making at a granular level it’s not there.”
Channel 4 is another organisation investing more in gathering first-party data, having developed a strategy that has resulted in the registration of over 11.5 million viewers in the UK. Head of data planning and analytics Sanjeevan Bala says first-party data “is both high quality and obtained through a relationship of trust. There are too many sources of third-party data that are increasingly perceived as lower value and often based on models, and therefore not as reliable for targeting.”
The Post Office’s real-time media decisions
Sunday 20 July 2014 saw the launch of the Post Office’s ‘We’re Changing’ campaign, which promoted that it has branches open on Sundays. The campaign ran across a number of channels, including social media, and the Post Office worked with Mindshare UK to use its Loop Room, a physical data hub that made statistics, campaign information and national news visible across six screens simultaneously, allowing informed decisions to be made based on the data at hand.
“We used external and one-to-one personal data to inform the degree of media spend behind certain social media posts,” says social media manager Darren Jones. ”We knew that by having all the right people in one room together we could make decisions faster and respond quicker to real-time events.”
The Post Office used Lithium, a back-end social media technology that enabled it to drill down into what individual users were talking about to give tailored responses. The organisation also responded to real-time events.
The campaign took place on Prince George’s birthday, so the Post Office created a bespoke post for Twitter, Facebook and Vine, and used targeting to make sure it was delivered to the right people, such as followers of the Royal family. The post read ‘First steps for Prince George, next steps for us’. The campaign achieved a 7.9% engagement rate, and prompted 300 tweets an hour at its peak.
“In terms of the reliability of social data you have to take into account context, and spend time interrogating it and making sure it is right by looking at past campaigns and activity,” says Jones. But he adds that ‘dark social’ – social sharing of content that occurs outside of what can be measured by web analytics programs, such as via online chat or email, adds another challenge. “You can interrogate all of this public data yet find you are sometimes missing two-thirds of the pie.”