‘What works can be measured, but not always in advance’ (Joy Howard, CMO Sonos)
Yesterday’s Trustees meeting of the Marketing Science Institute combined academic and industry experts on the ‘Why Behind the Buy’. As summarized by Carl Mela, people buy because of their preferences and their perception of your attributes. On preferences, Morewedge explained why people pay more for physical than for digital goods (sense of ownership), Lamberton contrasted time and quantity scarcity, von Nuremberg described consumer journeys and decisions rules by the trigger that started it, while Toubia and Grillo analyze preferences revealed through social listening and Goldfarb predicted that the lower cost of AI predictions will make other decision components more valuable, in particular judgement.
On attribute perception, Joy Howard talked about creative messaging, Fuster on social ads and Pauwels on how rival brand engagement increases consumer excitement and conversations helping your brand. My personal summary of the key insights from business speakers, in order of appearance.
First, Joy Howard (CMO of Sonos) contrasted Mad Men’s 3Ms (Myth, Magic and Meaning) with the Mad Scientist’s 3 Rs (Right Person in Right Place at the Right Time). Mad men (creatives) often fear big bad data because they come from a world where one failure can get you fired and because breakthrough ideas may be killed as they don’t test well before airing. Instead, data should be used to give courage to go against conventions, validate breakthrough ideas after airing, innovate and inspire: ‘what works can be measured, but not always in advance’. Sonos does not have A/B testing as a requirement, but does measure how people talk about the ad in earned media. When hiring analytics experts, look for humility and gentleness: they should show data insights when useful instead of getting so excited by it they don’t listen to anything else.
Next, Bank of America’s Paolo von Nuremberg showed that consumers’ checking account choices involve different triggers, decision rules and nonlinear journeys. Incorporating these different decision modes enables more effective marketing. We can segment consumers as either transitional (changing life circumstances), opportunistic (pursuing promotions), or aggrieved (upset by a fee or consumer service issue). Compensatory, lexicographic (go for the cheapest) and affect referral decision rules dominate for each respective segments.
Oracle’s Joe Fuster reminded us that TRUST is key in why consumers buy: he personally feels that with Amazon, Netflix and Spotify but not with many other companies. Once you have trust, you can personalize 3.0, e.g. targeting lactating women with Enfamil promotions as they opt-in when getting on Montreal public transportation. AI can also help your salesforce to answer questions such as ‘Who should I talk to now? What products have these prospects shown interest for in the past? What do we know about them? Which content should I use to make the sale?’ With GDPR coming, every business should think about what data they REALLY need and allow the consumer to decide up front what she shares with you and where you can come. For instance, if the consumer does not allow you into my house (Amazon) but only the trunk of her car, Volvo Sweden has a code that opens only your trunk so companies can deliver. Such AI solutions do not have to be expensive, eg chatbots can be text-only for millennials, we can embed an API in existing marketing tools.
Finally, Frank Grillo from Harte Hanks complained that technology has mostly allowed us to do bad marketing faster with e.g. creepy retargeting. Instead, marketing technology should involve genuine listening twice as much as talking to the consumer (we have 2 ears and 1 mouth). Harte Hanks maintains the principle of only interacting with a consumer that started the conversation – it does no outbound marketing at all, including paid search. Instead, it picks a larger conversation of relevance to consumers and publishes opinions about it. For instance, the piece ‘How brick and mortar retailers can beat the digital ones’ allowed it to inform consumers about its small volume logistics which helps retailers handle returns. When a consumer does give permission e.g. to email her after her fourth website visit, the company looks back at what she has done at the site before and asserts a few interest points – and then carefully tracks how the consumer responds. Some of this can be automated, but a human marketer is still needed for new situations and to formulate and test new assertions
So how can we understand consumers based on their psychology, behavior and interactions? Stay tuned for my next blog on the academic speakers at the MSI Trustees meeting!