We caught up with Nic Pietersma, Business Director UK for Ebiquity, ahead of his participation on our panel on metrics and finding the balance between short-term ROI & long-term brand success.
AAI: Why is it that more data can lead to fewer insights?
NP: I think with an abundance of data and increasingly powerful algorithms sometimes we forget the old tricks like the 80/20 rule. It is amazing how often the simple act of adding things up and ranking them can help us focus the mind on what really matters.
A related challenge is that practitioners in marketing analytics tend to oversell what they can do with data. Decision makers value confidence, clarity and optimism in their advisors and this leads to a tendency to overstate the accuracy of predictions and perhaps understate the width of the confidence intervals.
AAI: Where should analytic & insight teams sit within the business structure?
NP: That’s a really interesting question. Most of our client contacts sit within the marketing function, but ROI is so central to what we do at Ebiquity that we have always felt reporting lines into the CFO or Commercial Director could work well.
Probably more important than the reporting line is the mandate given to insight and analytics people. Personally, I always feel a bit uncomfortable when I see “making the case for marketing” spelt out as a bullet point in project goals. I think in our field identifying failure, and honestly identifying what does not work is just as important as identifying does work.
Organisations that recognise this tend to make better decisions, but sometimes this requires a cultural shift.
AAI: Why is it important to break down the silos between data teams and insight teams?
NP: Sometimes silos exist for arbitrary reasons; if data teams and insight teams sit in a very different place on the organogram collaborative work and coordination tends to be much harder.
Another barrier is lack of cross-functional skills. In a perfect world insight teams would have better coding and data management skills and data teams, on the other hand, would have more of an appreciation for how consumer and market trends relate to business objectives.
AAI: As an industry are we losing the art of interpreting human motivation?
NP: I’m not sure how to answer this question. What yardstick should we use?
AAI: GDPR should mean cleaner better data – has it?
NP: I think it is too early to say. Most of the measurement strategies we use at Ebiquity, such as econometrics and geotesting, do not rely on personally identifiable data. So the impact on our business has been limited.
AAI: If AI and machine learning could deliver one improved solution to help your understanding and targeting of your audience, what would it be?
NP: This really does depend on your business model. For some businesses getting to the right consumer at the right time and the right place is all important. But for most businesses, there isn’t really such a thing as ‘wastage’, at least not in the purest sense of the word. Target audiences are usually broader and more fluid than we recognise in the marketing community. Usually, cost-effective reach is more important than, for example, using AI to target slightly more accurately with some kind of dynamically-served tailored creative.
AI has the potential to be phenomenally disruptive in our economy; self-driving cars, logistics, call-centres, agriculture and so on. I think the impact of AI on targetted advertising is really more of a footnote.
AAI: Blockchain, could it be the answer to all advertisers’ problems, 100% accurate behavioural data or just the next new shiny thing?
NP: ‘New shiny thing’.
Blockchain, which I understand is essentially a decentralised anonymous ledger, often seems to be touted as a solution to problems that are already solved quite nicely by old-fashioned, centralised, non-anonymous ledgers… you know, like the way credit cards work, which are famously inconvenient.
I have read that blockchain may be a useful technology for digital ad-verification and that certainly is a problem in search of a solution. So for this, it is worth keeping an open mind.
AAI: What’s the biggest single challenge that impedes the successful integration of behavioural and survey data?
NP: It is unlikely that survey data and behavioural data will ever be perfectly aligned, but from a decision-making point of view it makes sense to have both reference points available.
Netflix made a change to their rating algorithm recently that is instructive. In the past they had a five-star rating system, but realised that people would kind of slip into ‘film critic’ mode when asked to rate a movie that way. What they really wanted to train their algorithm on was unguarded feedback on whether people enjoyed the movie or not.
Their new system is simply a thumbs-up or thumbs-down. It is more user-friendly and seems to work well… and arguably this is a small change that more closely aligns real world behaviour with the ‘survey response’ data.
AAI: Why is it important for brands and their agencies to be able to compare online and offline audience and advertising data more effectively?
NP: If a consumer sees the same 30 second advert on an iPad, through a Roku or smart TV, or on traditional channels it is a fair question for advertisers to ask what the relative costs are and where they would get the most cost-effective reach?
The question is complicated by differences in reporting standards and the definition of audience metrics – so far the industry has not done enough to make audience metrics transparent and comparable.
AAI: Can you sum up the holy grail of total advertising attribution in one sentence?
NP: No, we should reject holy grails!
Instead, the rally call should be for a diversified portfolio of research methods and a cultural commitment to openness.