Falling for data

In our Q&A, industry experts share advice for a healthy relationship with data and discuss key topics they feel need more attention. Q&A participants:

Marc Binkley (MB), MD & Digital Strategy Lead, Anstice Communications
Sonia Carreno (SC), President, IAB Canada
Sophie DeLadurantaye (SD), Director, Data Services Commercial Mail, Canada Post
Bryan Saunders (BS), CEO, Psychology and Marketing Inc.

Are terms like data-driven reflecting an unhealthy relationship with data?

SC: From the start, terms like this were rooted in the very promise of digital advertising. For the most part, they have encouraged investment in a highly measurable medium. However, we’ve run into some confusion because we haven’t consistently defined the data that matters. We also need to be more sophisticated in the way we use data to drive positive outcomes.

MB: No, but data is only valuable if it allows you to make better decisions. As marketers, we need to be more selective of the data we have. Not all data is good data, so it’s important to understand the source and quality before we start using it to make decisions. Otherwise, we might make good data-driven decisions about the wrong things.

BS: To the contrary: Even organizations that are leaders in embracing change are quite balanced, and hesitant to let data completely drive all their decisions. What we are finding is that data is causing more managers and business owners to have their long-held opinions challenged, which is leading to a lot of bold and innovative decision making.

SD: As the marketing landscape evolves, being data driven will become increasingly essential. However, using data alone to drive business decisions does not guarantee success. We can’t forget the role that humans play in turning data into valuable insights that are aligned with business goals. They say that if you torture data it will confess anything! To get the right insights, you have to start by knowing your market and your business. Then you need to ask the right questions.

How can data serve people better and create more value?

MB: I’ve two thoughts on this. First, we should view this from the customer’s point of view. Does the data help us plan new or better solutions to solve needs and pains? Secondly, we should ask ourselves, “Do we know that for sure or do we only think we know?” We can only see what’s available to us, so it’s important to consider that some data may be missing.

SC: Much like any science, the most surprising and important results are often obtained when we ask critical questions. For years, the industry has been satisfied with one-dimensional data sets that provide very narrow insights. When we ask questions like, “How much engagement will result in an investment?” and, “What kind of engagement might result in incremental investment?” we open the door to experimenting with more powerful creative and we better integrate our efforts within an organization.

SD: Every data point tells a personal story about an individual and their brand interactions. Humanizing the data reminds us to protect it as if it were our personal information. Marketers are privileged to be part of their customers’ lives, and relationships must be based on an exchange of value. By allowing us to collect data, customers leave us signals. Because every journey is unique, it’s up to us to figure out what those signals mean.

BS: Context is important. What’s the weather? What is happening in your city? Most importantly, what are your staff hearing from customers and what is their instinct on what’s happening? One way we’ve used data is to move away from industry jargon and match our language to that of our customers. If customers use completely different words when looking for a product, you need to adjust.

How can data increase marketing effectiveness?

SD: Data is at the centre of marketing effectiveness. It helps find the right customer and deliver the right message at the right time via the right channels. The cycle is completed when we measure what matters. In today’s omni-channel ecosystem we need to understand and attribute results to each channel in the media mix. To drive true marketing effectiveness, a test-and-learn approach is also key. If every campaign teaches us one new thing, it means we can make the next one more effective. That’s really the key to marketing success.

BS: We use data to better understand different customer types. Which messages are important to which customers at which times? It is no longer okay to say, “Half the money I spend on advertising is wasted; the trouble is I don’t know which half.” We’re always testing which creative and which ad-buys are driving the most value. Surprisingly, the best revenue generation tools often aren’t the best looking or the most expensive.

MB: I’m an advocate of measuring lift on business drivers rather than media metrics. If we can measure and report marketing impact on 1. Number of new customers 2. Average transaction value 3. Frequency/recency of purchase 4. Churn/lifetime value then we’ll start improving the effectiveness of marketing investments within the business.

SC: While using data to target advertising seems like an obvious best practice, it certainly doesn’t end there. There are many ways to increase marketing effectiveness with data. For example, you could leverage first-party data to enhance story-telling capabilities and guide content that resonates. You might leverage online-sales data to decide whether to expand into strategically placed pop-up stores. The possibilities are endless. From a policy POV, using data to understand risks and compliance responsibilities is critical (e.g. understanding the traffic generated on a property that originates from California will inform whether you must comply with the California Consumer Privacy Act). With global legislation growing in our sector, these instances are on the rise.

Read more from our INCITE blog series

What data traps should marketers watch out for?

BS: Confirmation bias and insufficient statistical power. Most business people haven’t studied scientific design or statistics – so data is often misused. You should be trying to disprove your hypothesis – not just looking for data that confirms it. Decisions are too often made based on a tiny data set or a split test that hasn’t run long enough. The results are not statistically significant: it’s bad data. The result? Garbage in, garbage out.

MB: I’ll refer to the science nerd inside me for this one. False positives or false negatives are the biggest data traps. Even with the most rigorous test methodology, we could get a result that’s not real. To avoid that, we need to think about running tests several times to validate the findings. Once we get the same result several times, we can start thinking the insight might be real.

SD: It’s dangerous to focus narrowly on a set of short-term metrics, like click-throughs and conversions. While these help drive real-time campaign optimization, relying on them too heavily can impact long-term success. Marketers should emphasize more big-picture metrics, which provide strategic insight and broader context [like comparing acquisition costs to customer lifetime value]. For innovative insights, leave your comfort zone and expand your go-to metrics. Focusing on the wrong metrics tells the wrong story and may lead to the wrong business decision.

SC: Attribution modelling is extremely difficult in a dynamic digital environment. Another area of concern is the epidemic short-term perspective on performance. Once a campaign goes live, the race for instant results is on and the realities of normal purchase cycles, as well as other market factors, are often not taken into consideration. Data is bigger than campaign reporting.