Ben Rickard is a Digital Marketing Consultant in the UK. He is the former Chief Digital and Data Officer at Mediacom. Rickard took some time to talk with Audiens about the lack of diversity in marketing, what’s behind the in-housing trend and how CDPs will evolve.
Audiens: In your role as a digital marketer over the years, what has been the most significant change/shift you have seen in your career?
Ben Rickard: The most obvious one is the significant shift towards adtech over the last decade. The underlying question is whether this new technology is actually delivering better work and better results versus what was happening before with more manually crafted campaigns.
In theory, we all understood adtech was an enhancement. However, we had to come to grips with continuous innovation and the new capability it offered alongside the data privacy issues and that came with it. The pace of New York and Silicon Valley moved so fast; the dollars flowed so quickly, it meant in most cases, people just accepted everything was okay. The industry, on the whole, trusted other peoples due diligence; it was an unspoken circle of trust. I’m not sure that is a healthy way to operate.
Audiens: Brands are beginning to move their marketing in-house, why has that become the trend versus outsourcing that to a digital agency?
Rickard: In the UK and Europe, we are in the thick of it right now. I believe we are coming to the end of the first cycle of this trend where a number of big brands have tried it and are beginning to see some initial learnings from it. Some things have gone well, and some things not so well!
On the whole in-housing Search seems to have worked quite well as it’s a contained system within a contained platform (Google) so brands have had reasonable success here. On the other hand, many brands have struggled with in-housing programmatic, as it’s much more complicated to operate and needs multiple partners and platforms to do it properly – it’s not as simple as jumping into the Google stack.
Talent retention is another big factor. Cool brands in good locations with nice offices are having much more success retaining in-house talent than others – after all who wants to climb a career ladder in a quiet industrial estate out of town? I know of several brands that have hired good talent to build out their in-house team but have then been held to ransom by those key staff when they decide they want to move on – sometimes negotiating a +50% raise in salary to stay!
Audiens: What are brands looking for today in terms of engagement?
Rickard: Brands need to start moving towards a ‘publishing model’ where they use their data to understand what customers want and then create more of it – this applies to advertisements as well as content. This is a big shift from the more comfortable ‘broadcast model’ and something brands have never had to think about before. Many brands have good customer data, but to date they’ve mainly used it for targeting rather than content creation. Brands must think like newspapers and obsess about what people are reading, what they are sharing and what they are buying, and then create campaigns/content off the back of that – not just pushing out what the creative agency has made for them! A good publisher will work out which articles are working well and which aren’t, and then change the headlines and images to make the article more popular – brands need to start behaving like this.
Audiens: How will CDP evolve over the next several years? Do you see it being replaced by another platform?
Rickard: In the next 18 months we will see that the vastly oversold DMPs will start to disappear; I know several high profile brands that are writing off these investments. We will also see multi-touch attribution disappear in the short term as we lose third-party cookies. The ability to understand and track a user’s movements across different domains and platforms will largely disappear, so attributing messages beyond last click will become very difficult. DSPs will continue, but they will be used in a more defined way within data ecosystems and walled gardens – third-party data to augment open web programmatic will add less and less value and probably disappear as well.
I believe CDPs now become a critical technology to any marketer that is serious about connecting their owned and paid messages as it turbocharges their first-party data and builds outwards towards paid messages – this is now the new model for marketing.
Audiens: What has changed for brands tech-wise in how they reach their audience?
Rickard: Brands are collecting more data and also have access to more processing power via cloud technology – this allows marketers to switch on new machine learning capabilities. This, in turn, opens up the fascinating area of ‘uncommon insight’ – things that are only seen by machines e.g. audiences that look different but act the same before purchase. Tribes will become much more important than traditional segmentation.
As we swing back to contextual targeting and think about the capability of machine learning, several big DSPs and programmatic players have also started to make changes to how their tech works. Instead of focusing their AI guns on better audience buying they are now building out machine learning models to improve contextual buying using signals from all the available non-personal data.
Natural language processing that understands the text in an article or image recognition on web pages suddenly becomes incredibly useful data points for contextual targeting. The early tests of this are very encouraging and show better results than the legacy audience targeting of the last decade. Adtech vendors are now putting a lot of R&D investment in this space because they know the industry needs to shift this way to remain privacy compliant.
Audiens: How will in-house teams use their data to create and train machine learning models?
Rickard: This is where the new gold is, and I have seen brands starting to do this very well. In-house teams are just beginning to build out their own ML environments, specifically for marketing teams. The main problem I have seen is where brands have lots of quality data in the business, but it’s locked down in legacy IT systems and difficult to extract. Connecting cloud systems with ‘off the shelf’ ML is starting to breakthrough here e.g. Google Marketing Platform now connects directly into Google Cloud Platform – this opens up some incredible capabilities for marketing teams without the need for IT support.
A great example is an insurance company that used their first-party data to build ML models of potential applicants likely to crash a car in a policy – much of this is based on non-personal data signals. When the insurance company sees an applicant who is considered a ‘high risk’ they remove themselves from showing ads or trying to acquire this new customer. Conversely when a high value, ‘low risk’ applicant surfaces, they go out their way to outbid the competition on digital media. It’s early days here but those brands who are doing this now will simply pull away from the competition within the next two years – data transformation and AI must start now or you will not survive.
Audiens: If you could change anything in the digital industry, what would that be?
Rickard: Diversity. We just don’t have it, and we need to change that. The more variety of data you have in a model, the better the output. The same applies to our industry, you get better work when you have different inputs from different kinds of people.