Last week, I spent a day at iMedia’s Data-Fuelled Marketing Summit. I turn up at any conference with the word ‘data’ in the title with some trepidation, but, thankfully, this was an event that was free of data hyperbole. There was a refreshing focus on the broader context in which data is applied to marketing, how brands are practically making use of data to improve user experience and drive marketing performance, and challenges to the orthodoxy of marketing being defined by data science.
Here are 10 useful things I took away from the day:
- Data is about people. Many speakers on the day, including our own Paul Frampton and Marc Mathieu of Unilever, emphasised the importance of customers in data-driven marketing. In other words, data exists not just to target people better, but to give them better experiences. If we don’t do this, customers will increasingly look for ways to take their data back from brands.
- Use data to improve customer journeys, not just acquisition. If that is a starting principle, it requires re-thinking how we design customer journeys underpinned by data. Chris Duncan of News UK spoke about enchanting – not stalking – customers in the journey from prospect, to subscription, to readers getting the most from Times+. He said hard promotional tactics were the least effective at driving subscriptions with good lifetime value. Sometimes, the ‘next best action’ is to suggest an article or another way to get Times+ content.
- Brand strength enables data-driven marketing. Data-driven performance media cannot be planned, managed or measured in isolation from brand-building. Roland Bryan of Wowcher warned against over-emphasising short-term efficiency over long-term brand and commercial value. Wowcher has found that acquisition efficiency and long-term ROI (again measured against lifetime value vs. pure acquisition) are driven by a response to a brand positioning – a differentiating brand proposition, delivered effectively, allows acquisition and marketing economics to work.
- Data should drive de-commodification. The application of data will inevitably get more complex as computing becomes more ubiquitous across devices and in our daily lives. The long-term objective of data in marketing, according to Dr. Brad Berens, is not about managing or quantifying user journeys. It’s about building a relationship with customers that creates switching costs or cognitive load for those customers who may be tempted to switch. This is really about how using data creates a service for customers brands don’t want to lose eg Nike creating Fuel as a currency for exercise.
- Less is more. Data-driven marketing enables us to measure many variables, but there’s a danger this can lead to stasis. Data gathering and analysis should always ladder up to a core set of critical commercial KPIs. Establishing these, and making sure they are accurately measured, should be a strategic exercise. Wowcher has two critical commercial KPIs and five further KPIs which ladder these up – seven is its maximum.
- Cross-channel attribution is (really) starting to arrive. It may be early days, but attribution is moving beyond single clickstream models and into models which blend digital and ‘analogue’ datasets. There isn’t a silver bullet here, but brands are building potentially different (and complimentary) approaches eg real-time econometric models, advanced data warehousing (ie single customer view), universal analytics.
- The coming of the data warehouse. The DMP is an industry unicorn, often spoken about in mythical tones but rarely seen in action. It looks like 2015 could be the year that we turn talk into action, building DMP solutions in whatever form (eg bespoke builds, rollout of DMP solutions from the likes of Google, or specialist third parties such as Teradata) that start to build more universal customer IDs tying up intelligence and action across touchpoints.
- Quantifying the pre-sale. Whether DMP or attribution, there was an emphasis on figuring out, and best accounting for, the relationship between upper-funnel marketing and customer life-time value. Teradata is building models and solutions that focus on pre-sale activities – such lead generation or pre-registration – and finding that nurturing the pre-sale relationship improves both conversion and lifetime ROI.
- Behavioural-driven media planning is coming. The growth of addressable media will start to change the dynamics of media planning and measurement. Dunnhumby is starting projects that will fuse its data with channel owners (notably Sky Adsmart), with the intention of using these fusions to inform channel planning, optimisation and measurement. This will be done against actual people and their purchasing behaviour, not just audience segments.
- Programmatic – the need to educate and dispel myths. There was a session dedicated to transparency and trust, attended by both clients and agencies. One thing is clear: agencies (or industry bodies) need to do more to lift the veil on programmatic, as well as why and when brands should be using it, how it best works – taking it way beyond a line on a spreadsheet. Clients clearly have lots of lingering questions about programmatic, and are increasingly likely to test different solutions (agency desks, specialist third parties, big tech vendors) in parallel. Encouragingly, clients see the value that agencies bring to programmatic: GfK argued that access to quality inventory, and private marketplaces, are key to driving performance – and agencies have helped unlock that for them.