Speaking from the floor of this week’s Dmexco 2015, Laura Desmond, Starcom Mediavest’s CEO, speaks to The Drum about ‘the new brew’ of using data to inform a content strategy, how the future of media-mix modelling is through attribution, and how consumers’ shortening path to purchase will challenge retailers.
This week’s Dmexco conference runs with the tag-line ‘Bridging Worlds’ to discuss the disruption facing the marketing industry given the risen of technology, and data-driven digital advertising techniques. Here, Starcom Mediavest chief executive, Laura Desmond shares insights on how her outfit is adapting its approach to content production, media-mix modelling, and how the shortening of the purchase journey will disrupt the dynamic between retailers, and FMCG outfits.
The Drum: How are you using data to inform content?
LD: With our client Heineken, we are doing tremendous content programmes, with partners like Twitter and Google, and we can do a mix with them through real-time, and social in order to tell stories.
Heineken has really embraced the ‘new brew’ with content and data. Data can inform the content, with content then inspiring the data you get. So from this you can see how the two come together, and create relevancy.
The results have been positive. For example, with our work with Twitter around the [FIFA] World Cup that’s a great example of how we were getting results in real-time on how many people were engaging with the campaign. We were then able to track those people through a combination of attribution using third-party data to link them to sales growth over the course of the campaign.
Through the attribution modelling, we were able to show very clear business-lift.
So with real-time data, you are able to adapt and adjust [a campaign], and then with attribution modelling, you are able to show how the programme worked, and whether or not the activity helped seal the deal on a path to purchase. For FMCGs that’s what you need.
We can get real-time dashboard results on an hourly, and daily basis through digital partners, the attribution modelling takes a bit more time. The future of content is adaptive, and in short micro-cycles.
The Drum: So how has Starcom Mediavest adapted to the decreasing ‘news cycle’ to make your clients’ content as relevant as possible?
LD: About 18 months ago we acquired a company called Relevant 24 – a real-time content company – and they work to turn around content in 12 or 24 hour cycles. They use proprietary software that scrapes the web for the most popular things people are talking about every morning, and then match those moments to brands that are in a campaign. They then take about four-to-five hours to produce video and display ads that are aligned to moments that we can associate a brand to trending topics.
They then put them up as paid media, or introduce them to the earned channels to gain impressions, and then the next morning, we’ll do the same again.
The reason we acquired that company, is that we think the future of content is adaptive, and in short micro-cycles. A great example of that is with Coca-Cola, for the 2004 Olympics 10 separate 30 and 60 second ads for the Olympics, last year in Rio [sic] they produced 10,000 pieces of content. So in 10 years, that’s the step-change.
A real-time 72-hour cycle is a product we developed called content-at-scale. This is a product that assembles pre-approved copy assets on a tech platform, and targets them against pre-determined audiences based on their behaviour on the web. But it does this algorithmically, and then serving it in real-time as native ads. We are using that as a substitute for display ads, and we can do that in real-time, on a day-to-day basis, or every three days.
The Drum: So how are you able to tell this strategy actually returns an improved ROI for a client?
LD: One of the things we learnt with a client like Walmart when working with partners like Facebook is that there are some blunt metrics to show that things are working. For instance, if we put an ad in the Newsfeed saying that ‘bananas are 99 cents only today’, that within 15 minutes, they are flying off the shelf’.
However, then we have the attribution modelling that is a full suite of modelling that helps you understand the full path to purchase, which is what we’ll do to ‘officialise’ the results that we see in real-time from an engagement, and retail standpoint.
We think attribution is the future of mix-modelling. You have to understand attribution that gets you to the path-to-purchase; mix-modelling doesn’t get you to the total story.
The Drum: What are the practicalities of resolving an attribution headaches for a client?
LD: We have a proprietary multi-touch attribution (MTA) model that we work with our data scientists and researchers to improve. We are easily doing MTA for about a third of our clients. We’ve gone head-to-head with MTA players like Google, Adobe, [Oracle’s] Blue Kai, and a few others, and we have won our fair share of independent work for our own clients. So they stand up with the best-in-breed.
How you do so depends on the client. The reason being because the path-to-purchase for every brand is very different. The media-mix is different, the purchase funnel is different. There are some general trends though.
For instance, social and TV are a very powerful combination. Out-of-home (OOH) with mobile, is powerful, especially for retailers. Mobile video is very powerful for connecting with millennials.
The Drum: So how is the blurring of digital and ‘real-world’ behaviours among consumers challenging clients?
LD: We’re starting to see the purchase journey shorten, and collapse into the media journey with the product development of companies like Facebook and Google, with them aggressively growing brand and retail partnerships around ‘buy now’ buttons on their platforms. This means that a consumers’ media journey can also be a commerce journey online. This is going to be incredibly disruptive to how retailers sell in both the physical and virtual world.
Read more from Dmexco in The Drum's dedicated content hub.