This week, the subject is data-driven marketing. Click on the speakers’ names below for an overview of their session.
What do you see as the biggest advance in data-driven marketing so far this year?
Nakisa Mavaei, The Specialist Works: I’ve found a real drive for ‘traditional’ publishers to offer campaigns that can target consumers by interest as well as behaviour. This has led to some very innovative campaigns and publishers utilising their customer data in an experimental way.
Matt Wheeler, Driftrock: With the launch of LinkedIn Matched Audiences, it's now possible to reach customers from your database across all the major platforms; Facebook, Instagram, Google, Snap, Twitter and LinkedIn. Customer data just became 10X more powerful.
Tom Manning, Forward3D: We feel the biggest advance was a relatively quiet announcement that YouTube, and then other Google services were moving away from cookies to track users across properties and devices by mapping other elements such as e-mail. Cookies currently have two major drawbacks; their lack of suitability for cross-device, and their unfair general perception of being 'bad' for users. The move away from a reliance on cookie based tracking, and diversifying potential tracking methods is a major step towards a more accurate deterministic cross-platform tracking solution.
Stephen Pavlovich, Conversion.com: Within conversion optimisation, it's the increasing sophistication and prevalence of experimentation platforms. Google launched Optimize, their new 'freeish' testing platform. Meanwhile, existing products like Optimizely have added significant functionality with Full Stack for server-side testing.
What’s your vision of ‘perfect’ data-driven marketing in the realms of what’s actually possible with today’s level of technology?
NM: In an ideal world, it’d be great if we could match the consumers' interests and profile with that of the publishers in a truly automated way. For the time being, we’re going to have to get around a table and discuss what’s possible. We’re only going to get there by being upfront with who buys what and who interacts with what.
MW: Connect all your customer, online and offline channel data together to a single view. When you have a clear picture of that person and where they are in the lifecycle, you can serve the perfect next message, while tracking everything.
TM: In our opinion, 'perfect' is impossible to achieve as there will always be gaps and discrepancies in tracking user journeys that can never be completely eliminated. However, marketers now do have all the tools available to be able to understand the impact of each campaign they run. Advanced statistical modelling techniques and the aforementioned improvements to tracking technology enable activity that drives positive results to be more easily identified than in the past. This can lead to better budgeting decisions that allow advertising to more efficiently achieve business KPIs, as well as more targeted campaigns that resonate better with the relevant demographics. Our vision, therefore, understands the limitations of existing measurement and fill in the gaps with statistical modelling.
SP: Perfect data-driven marketing requires three key components: the ability to track the full customer journey, to analyse and understand behaviour and then to experiment and iterate. All of is possible – and when applied to a website with significant traffic, it allows companies to scale rapidly.
What’s stopping us from getting there?
NM: I feel a lot of it is down to transparency and issues with potentially getting those audience profiles. I also feel the publishers need to continue to offer campaigns which will continue the development of relevant and targeted campaigns.
MW: It's hard to connect the online and offline data into a usable format for marketing then deploy it; disclaimer - this is what we do at Driftrock!
TM: There are a number of legacy issues which make this transition tough for some advertisers. Whilst the industry, in general, is moving towards data-driven attribution and other more insightful tracking and measurement techniques, understanding of these methodologies outside of marketing departments can be limited. This means that it can be challenging to get the wider business to buy into the insights and the actions that should result from them. Also, in many organisations, budget is still siloed by channel, which can prevent it from being deployed in areas where it would be most effective for the business.
SP: The blockers are frequently awareness and technology: either not understanding the importance of analysing and optimising the user journey, or not having the tools or ability to do so - especially when the user journey veers offline.
Without giving too much away, how are you approaching the subject of data-driven marketing at PI LIVE?
NM: Through discussion. I feel we as marketers need to understand the audience of the brand as well as that of the publisher. Only through being open can you really start to get creative and really see what’s possible.
MW: I’ll be explaining why people-data makes all these strategies possible. It's the glue for true multi-device measurement, lifecycle marketing and predictive marketing with machine learning.
TM: We're discussing some of the topics above, particularly the need to accept that your data can never be complete and why. We will emphasise that it should not stop anyone from using the data they have to make correct decisions and that businesses can do more harm to themselves by aiming to reach the impossible state of 'data perfection' than by making careful decisions on incomplete or imperfect data.
SP: We see experimentation as a crucial step in data-driven marketing – without it, we're relying on - at worst - opinion and - at best - limited research and best practice. It's only when we segment, experiment and analyse that we can understand and modify user behaviour, and drive our companies to better performance.