Television works for advertisers. It works better than other media across the full marketing funnel – from building awareness to moving perception and eventually driving sales. Proving this in a scientific way has not been easy, and the attribution solutions that have come to market in recent years focus on sales.
But sales is only part of the picture. In fact, if you focus on sales, which is the outcome you get from the bottom of the classic ‘marketing funnel’ or ‘customer journey’, you miss much of the work that television does at the top of that funnel. There is still a danger that marketers end up under-valuing television. You also miss the chance to optimise campaigns to improve top-of-funnel metrics.
This sums up the thinking at 605, the data analytics specialist that has been helping major brands like Walmart and Uber to understand how television shifts attitudes and behaviours at all levels of the funnel. 605, which focuses on the media and entertainment market, has been attributing outcomes to advertising exposures but, just as importantly, this knowledge has then been used by clients to optimise campaigns moving forwards.
In fact, a key feature of 605’s work is incremental optimisation. Campaigns have been adjusted. Consumer attitudes have been tested. Campaigns have been further refined. Attitudes and sales uplift have been noted. Clients have made use of audience-based buying of classic linear TV, creating custom audience segments and then finding the networks, dayparts and shows where the target group can be found.
A programme of regular consumer surveys has created the benchmark against which attitudinal change can be judged. Survey data, viewing data, shopping data and other measures are directly linked to verified household details (in a privacy compliant manner) so there is a clear cause-and-effect logbook.
You can see which ads have the biggest impact on brand favourability. Marketers can identify the broadcast networks where the most ‘persuadable’ consumers watch their TV. You can measure the time-lag between improved brand favourability and incremental sales.
These capabilities, which have been used commercially for 18 months in the U.S., have now been productized by 605. The result is ‘The 605 Impact Index’, which was made generally available to brand clients and agencies last month on a managed service basis. Programmers can also license the service as something to resell to their clients. Initially this product is focused on the U.S. market but 605 is already looking at future international opportunities.
The solution draws on what is claimed to be the largest matchable TV dataset in the world – although currently limited to the U.S. 605 has relationships with MVPDs (Pay TV operators), one of them being Charter, to harness second-by second viewing data as the evidence for what households are watching and which advertising they have been exposed to.
The company is pioneering the model for federated data analytics in this marketplace with its multi-operator inputs to its DMP (Data Management Platform). The viewing data covers 12 million households today (a footprint that includes more than half of the top 20 advertising TV markets in the U.S.) and will rise to 15 million within the next few months (then covering every DMA – designated market area).
A data safe-haven ensures personally identifiable information is protected. Pay TV customer files (once anonymized) provide authenticated household credentials. As Ben Tatta, Co-Founder & President at 605, points out, “The underlying source of truth is the household file. You get the household composition – what people are in the home.”
605 views this census-level data as a superior base from which to build an attribution model, when compared to traditional TV ratings and sample-based methods.
The advertising exposure in these homes is linked to first-party consumer research that is anchored around regular surveys. The survey process itself has been streamlined and made more cost-effective, which allows 605 to move beyond occasional one-off studies on behalf of brands to a regular benchmarking programme. This delivers consistency and means brands can judge the outcome of campaigns and optimisations against an already understood baseline.
In marketing, the cost of surveys is partly determined by the proportion of consumers you can get to respond to your first outreach. Generally, consumers have to remember which TV channels and shows they watched before revealing whether they remember seeing a brand advertisement, like for Walmart. This can be burdensome. Tatta points out that marketers can leave out the ‘what did you watch’ portion of a survey if they already know the answers, thanks to the STB viewing data.
Meanwhile, 605 data scientists who specialise in surveys have been helping to design the questionnaires and make them simple. The result is that the surveys are as little as half the size of what consumers may be used to. They are easier to complete, and that means higher response rates. Higher response rates mean you hit your sample size target faster, with less money spent.
According to Ben Tatta, the responses are more accurate. There is less post-sample data modelling to iron out obvious inconsistencies. Blending big data and survey data has been known to take a long time – but with ‘The 605 Impact Index’ one of the key benefits is how quickly actionable insights are available.
605 has been ‘taking the temperature’ of consumers at least every month. “The whole idea is that we have an ongoing measurement. You have a stable model to project upon,” explains Tatta. He contrasts this approach with checking consumer attitudes occasionally, like when a campaign is in-flight, which gives you less guidance about trends.
Walmart has gone public about its use of what is now known as The 605 Impact Index to identify the relationship between television, brand reputation and consumer purchasing behaviour. “It has yielded tremendous results,” according to Aaron Bernstein, Senior Director of Insights and Advocacy at the retail giant.
Tatta provides some detail on how Walmart used this tool to optimise its television campaigns on a step-by-step basis. The ‘client’ was the group within Walmart responsible for reputational marketing so their key performance indicators (KPIs) were brand favourability and net promoter scores (NPS). They wanted to see how television advertising made audiences more responsive to their messages. They also wanted to understand the knock-on effect on sales when the brand outcomes changed.
Walmart identified various consumer groups they wanted to reach, including shoppers who gave Walmart a ‘low share of wallet’ but who were ‘persuadable’. Every month, 605 provided feedback on brand perception and favourability and this identified the shoppers who could, indeed, be persuaded to shop at Walmart. The next step was to work out the best advertising inventory for reaching them. Broadcast networks and dayparts were ranked by their expected effectiveness.
It was discovered that some networks and dayparts contained lots of the persuadable shoppers, yet were not on the planning schedule for the Walmart brand campaign. This was duly changed. Having closed in on their target, the marketers then investigated the impact of different creatives.
“They had three creatives and they all did a great job driving brand favourability, but there was one that did not do much for sales,” Tatta recalls. “On the subsequent campaign flight, two pieces of creative were used and the third was used for a study. So, there was optimisation on the creative that helped drive lift across both brand and sales.” Monthly CRM data meant the marketers knew when one of their persuadable shoppers made a purchase at Walmart.
As Tatta points out, a key benefit of full-funnel attribution, which closely tracks brand health metrics and not just sales, is that you can see movements in the top and middle of the funnel as they are happening, and how this correlates to sales. Walmart demonstrated how this knowledge becomes actionable. “You can incorporate different criteria that you want to evaluate each time,” Tatta says of the ‘campaign flight, optimize, campaign flight, optimise’ sequence.
Walmart’s use of what is now The 605 Impact Index is also a good case study for audience-based buying of classic linear TV. Having isolated the persuadable shoppers, the company optimised its standard linear plans, still buying the advertising in the ‘traditional’ manner. “With linear optimisation, you use data and custom audiences to better inform your choice of schedules,” Tatta adds.
Audience-based buying, like addressable TV, relies on a more granular understanding of who an advertiser wants to target and where to find them. It goes beyond segmenting the population into broad age brackets and gender and instead identifies life stage, lifestyle, attitudes, purchase intent, and more.
Some examples of the attributes that can be used for audience-based buying, taken from various sources, include first time car buyers, travel enthusiasts, luxury shoppers, people who are ‘in-market’ for a car purchase, people who buy competitor brands, etc.
Media houses offering addressable TV or audience-based buying can pick-and-mix the attributes a buyer needs to create a custom target audience. They must also be able to show where those kinds of people reside (in terms of the media they consume). 605 has 5,000 data attributes that it can draw upon to identify consumers, segment audiences and help buyers to target.
“Our 5,000 attributes can be appended to any given household,” Tatta confirms. He points out that the most popular segment for audience-based buying, at least in the U.S., is new-movers – the people who have just moved home and are on a spending spree to fit out or improve their new place. “That is the most common use case for running an optimised linear schedule or an addressable TV campaign,” he reveals.
Advertisers can use their own first-party data to build the audience segments, of course. Tatta says exporting segments from a client CRM (Customer Relationship Management) system is considered the Gold Standard in data-driven TV marketing.
In a very simple example of how audience-based buying works, if Walmart wanted to promote an organic detergent, it could target households that purchase lots of organic products. Then, working with 605, it could use the measurement and attribution process to work out which homes, in that group, are more responsive to TV advertising and then embark on the kind of optimisation process outlined above.
Brands could use the system to optimise frequency and even the sequencing of ads (i.e. which creative someone should see first, and which creative second, to maximise impact). You can gauge the impact of specific product promotions.
Tatta is a big fan of television as an advertising medium and believes one of the industry-wide benefits of the Impact Index is the way it demonstrates the all-round power of TV advertising. 605 has seen that sometimes you have to be patient if you want sales conversions, with these appearing only once brand favourability has been lifted – a process that can take months rather than weeks.
“We have found in our studies that people with a favourable sentiment towards a brand, but who give them a low share of wallet, can be ‘moved’ within weeks. But if you begin with low brand favourability, it can take a long time,” Tatta reveals.
The former President of Cablevision Media Sales (the sales division of Cablevision, one of the pioneers in data-driven audience planning and buying for TV), points out that if you judge your media based on conversions over too short a period of time, you miss positive effects that TV is producing and which will eventually play through into business outcomes. Walmart has been using the 605 attribution tools for over 18 months to demonstrate the longer-term impact of TV.
“Everyone’s gut instinct is that TV delivers. It has the biggest impact at the top of the funnel. It performs across the full funnel better than any other medium, but it has not been measured properly,” Tatta declares. He is concerned that, in efforts to compete with digital, the TV industry could put too much emphasis on sales attribution. “Even now, so much of the attribution is bottom-funnel based,” he says. “Our data demonstrates how TV influences everything in the funnel.”
The brand-focused metrics matter, and the 605 Impact Index shows if someone has moved from having a negative sentiment towards a brand to a positive sentiment, using net promoter scores, for example. The metrics are organised into two distinct categories, based on campaign objectives. The brand impact metrics include awareness, brand perception, favourability and NPS. The sales impact metrics include ad engagement, ad response, sales conversion and conversion lift. Sales data that is fed into the system can come from online and bricks-and-mortar engagements.
Tatta emphasises that the Impact Index is not about reporting, but about action. “Reporting simply presents the news back to brands. You might get a negative read on the brand, but it will not show you what to do, other than run more TV. This [the Impact Index] is prescriptive.”
The attribution modelling takes account of other influences that can lift brand or sales metrics, so revealing the incremental lift that comes from television advertising. Thanks to its partnership with Hive Media, 605 even understands the earned media, as well as other bought media, that is influencing consumers. These could include a clothes sponsorship in a sports tournament or a storefront logo that appears in a programme.
Tatta says it is still early days for media-mix modelling, which goes beyond TV optimisation and orchestrates planning across both television and digital to achieve maximum marketing impact. He reckons 605 is helping improve one of the two big data sets needed. This product could result in television taking a greater share of that mix, too. “TV is driving everything but is not getting the credit. I think television is under-valued,” he declares.