Analytic Partners’ President, Nancy Smith, explores the challenges marketers are facing with Multi-touch Attribution.
Bowflex, Shake Weight, ThighMaster, Tae Bo, Ab Roller, Richard Simmons Sweatin’ to the Oldies… no matter what your age, and even if you’re a zealous cord-cutter, I know you have seen or heard of some of these “As Seen on TV” fitness products. What they all have in common is The Promise: Buy this product and you’ll look great in just 6 weeks without any hard work or dieting, and all for just “3 easy payments”. But how many buyers have ever achieved thighs even remotely like Suzanne Somers’? Few. Which is about the same proportion of marketers who have realized great marketing effectiveness gains from Multi-touch Attribution (MTA).
I have a cousin who used his Bowflex as a clothes hanger for a number of years before he finally gave up hope on 6-pack abs and sold it on Craigslist. To this day, my Cousin Freddie (name changed to protect Frankie’s identity) will tell you adamantly that the Bowflex just doesn’t work. Now, we all know that resistance exercise works to improve strength, tone muscle, and improve overall fitness. In theory, Bowflex should have worked for Freddie just as the Ab Roller, ThighMaster and others should have worked for millions of out-of-shape consumers. So it wasn’t the product that was flawed, it was The Promise. In real life, resculpting your physique takes work. It is not fast, easy, nor fun, especially when you’re just beginning. Like many others, due to his poor personal experience with Bowflex, Freddie has sworn off of all exercise programs entirely, and I can tell you that is a huge mistake and a huge missed opportunity, and it has helped create a huge belly for Freddie (which, ironically, is the area where he wanted to achieve that elusive 6-pack).
Now, I see that same Freddie-like level of disillusionment from marketers who bought into MTA platforms promising fast results. Marketers are feeling like the promise hasn’t lived up to the hype. Industry groups have conducted surveys, designed meta-studies, and issued reports to find out why MTA is developing a reputation for failure and the need to define best practices.
Is it possible to get the analytical equivalent of 6-pack abs with MTA alone? Hollywood stars and models who get the best results from exercise hire personal trainers to assess their lifestyle and transformation goals, customize routines to their particular needs, coach them holistically on diet and keep them on track to achieving results. The same goes for marketers who are trying to achieve accurate and actionable attribution. While any one or few of the tools and techniques might seem familiar or “easy”, the combination and methodological rigor necessary to adapt the solution to your “marketing metabolism” most often make the difference between success and failure.
Let’s face it, when people decide to buy a Shake Weight, they buy into the idea that the device will make their biceps and shoulders look like the bodybuilder in the commercial. However, honest inquiry into the motivations of the buyers will reveal that the buyers want more than bulging biceps. Most people want to look good, to feel good, to increase stamina and capabilities, and to have the long-term benefits of good health. A single device, or exercise, is not going to provide the benefits that buyers are seeking. Indeed, exercise alone, without changing diet and lifestyle habits, will most likely result in disappointment. Likewise, MTA, in isolation, can only provide finite gains in the absence of a holistic approach and the change management required to improve a business over the long-term.
Just like you wouldn’t use a ThighMaster to tone your triceps, MTA works really well in specific use-cases where excellent data exists and the walled gardens of Facebook, Google and others don’t limit the full 360 customer view. Used too broadly, MTA produces unreliable results. Properly applied, MTA is most often combined with a top-down view using econometric techniques or Marketing Mix Modeling and randomized tests. Three example use cases from our own experience demonstrate the point:
|USE CASE 1||USE CASE 2||USE CASE 3|
|Industry||Retail||Financial Services||Consumer Goods|
|Business Goal||Drive topline sales||Drive customer acquisition and retention||Grow the stagnant category, improve branding and market share|
|Data||Robust store selling data, limited customer and e-Commerce data||Robust customer level data, integrated online & offline activity and sales||Syndicated sales data, household panel, brand equity, in –depth capture of online & offline marketing|
|Key Performance Indicators||Daily & seasonal sales particularly in season||New customer acquisition, churn rate||Category growth, market share, Brand Equity|
|Adaptive Solution||Mix with MTA: Marketing Mix for holistic business view, MTA for digital deep dive (given data limitations focused on website store locator visits)||Integrated Mix & MTA: Integrated Marketing Mix and MTA for new acquisitions, Independent Churn Model, Customer Journey Analysis||Segmentation, Mix with MTA: Buyer segmentation, cross-brand Marketing Mix, Econometric model for brand equity, MTA for digital deep dive|
|Methodologies||Hierarchical Bayesian, Probabilistic discrete choice with ML, Multi-level optimization, Store and market tests||SEM model for direct and indirect effects, Hidden Markov, Multi-objective optimization, Experimental design||Latent class segmentation, Attraction model, Probabilistic discrete choice, War-gaming optimization, Market tests|
|Results||34% ROI improvement on $83 million in marketing spend||+12% in new customers, 5% reduction in churn rate||Identified segments & marketing to spur category growth, initial results show positive momentum|
While all cases leverage econometric and probabilistic models as well as all-important test design for in-market validation and optimization algorithms, the choice and mixture of these techniques are adapted to best meet the business goals.
Every business is different, and every industry has its own challenges. What works for one business will not necessarily work for others. The rapidly changing and complex multi-touch world of marketing as well as the outside competitive, economic and market forces necessitate a more robust and adaptive methodology – to address both the challenges of today’s marketplace and the unknowns of tomorrow. Multi-touch attribution is a big part of that future-state and one of the many tools that you must leverage correctly to address your specific business needs. Just don’t expect to see results in 6 weeks without working up a sweat.
For more information and industry perspective on MTA and overall analytic solutions, please visit our website (www.analyticpartners.com).