I recently read an article in The Economist’s Schumpeter business and management column titled, “Little things that mean a lot.” The article’s key takeaway is that, based on all the hype around big data, managers across all industries have been misled into thinking big data approaches would give them “massive, instant, Holy Grail solutions.” But in reality, most big data victories come in small advances. Schumpeter argues that this may be the unspoken secret of big data – marginal changes, when multiplied by a huge number of instances, or allowed to work over a long period of time, can produce significant effects. Or, as Schumpeter quotes a nameless Facebook scientist as saying, “Given the massive scale, even small effects can have large aggregated consequences.”
This is 100% true, and I see it all the time with the retailers we work with. Retailers can have massive scale, both in terms of number of SKUs and customers, and massive data (years and years of detailed transactions, weblogs, etc.). Their challenge is how to get the advantages of that scale, but still be relevant and personal with every consumer. We see this challenge particularly in email marketing.
Unlike direct mail, where you worry about picking the audience (because it is too expensive to send to everyone), in email the marginal cost to send to your full list is essentially zero. And so most retailers do just that. This approach, however, risks being irrelevant to too many customers. The conventional wisdom to counter this concern is that email marketers should do more segmentation. Abandoned cart, highest spenders, new customers, bought in last 90 days, etc. These segmented campaigns do work, in that they can have higher metrics (X% higher click-through, for example), which is great, but each segment generally does not impact enough customers to really move the needle from an absolute dollar benefit (what % of your customer base abandoned the cart in the last week? 1%, maybe?). Nor does it generally justify the added operational costs and complexity from running longer-term segmented email strategies. This is why “batch and blast” remains the dominant marketing strategy in retail today.
Enter big data. By leveraging the massive amount of data retailers have – coupled with the right predictive science and execution technology – retailers ARE now able to be more relevant to every single customer on their list. Note that I said “more relevant.” I didn’t say “perfectly relevant.” “More” means that you can use big data to become marginally smarter about all your customers so that for any one customer, your chances of getting it right are higher than before. Instead of headline improvements like “300% increase in revenue!” across a small segment, this approach creates 10-20% improvements, but across the full scale of your customer base. Given that email marketing is often the #1 or #2 most important revenue-driving channel, 10-20% improvements across the full list moves the needle! As Schumpeter quotes in the column, “It is about building a mountain with pebbles.” Small wins with big data add up.
Schumpeter pushes this thinking further by attacking the well-established 80-20 rule. This rule states that you can get 80% of the benefit by focusing on the most impactful 20% of your customers. But what Schumpeter argues – and he/she is exactly right – is that the 80-20 rule isn’t how business has to behave in a world with big data. Big data lets you efficiently impact the full 100% of customers. This certainly can and should improve the 20%, but if you can also marginally improve the performance of the other 80% at the same time with no additional effort, then that can create huge leverage for your business. It is a very similar argument that Chris Anderson famously termed as “long tail economics.” And it is exactly the dynamic that we see with our retail customers.
Improving your marketing success through big data isn’t going to happen overnight, but it is possible to make marginal improvements to become more relevant to your customers, which will certainly impact results. It will take time to prove to your customers that the content you deliver is uniquely tailored to them and is worth looking at. But once you get there, your marketing will move from being a “blast” selling tool to a trusted resource for each of your customers, and the foundation for a long term, mutually beneficial relationship.
How is your organization using and measuring the effects of big data? Are you only looking at big wins, or are you recognizing the small gains being made? Comment below or connect with us on Twitter.