How social media killed traditional research

If All the President’s Men was made today in place of 1976, the phrase “follow the data” would have been used rather than “follow the money”. Data miners were already living more lavishly than gold miners; now they have the power to play God every now and then. This isn’t an article of caution though, this is about how for marketers, social research this post-Covid-19 decade would become the only credible source of information upon which business decisions shall be forged. The influence in consumer decisions are omni social and the expectations of audiences are to be understood without being disturbed.

Unprompted intimacy vs prompted structure

The greatest boon of social media research is its possibility to seamlessly find insights without asking a single question. Accessing the unprompted, intimate feelings and opinions through social listening and audience intelligence tools. In my class, I explain this with an analogy of researching a tiger in a zoo versus in its natural jungle habitat. When prompted by a third person, there shall always be the chance of the respondent replying in societal pressure. How many men, for example, would accept that they would love to play Candy Crush over Call of Duty or rom coms is their favourite movie genre when asked on a survey? Yet, for Amazon, Google, and Netflix, it is an effortless method to precisely know this from their organic behavioural pattern. How many rom coms did they watch over action thrillers in the last quarter or which game on the Play Store did they invest maximum time on. This research lives up to the true sense of the phrase—the data never lies. Here data is mined through actions taken over the old school method of believing the promised words.

Data points depth

Facebook reportedly has more than 52,000 unique attributes classifying each one of us. With more than 29,000 lists of categories which it publicly provides to its ad buyers. Take a moment to seep in this Black Mirror-ish sentence. And answer one question honestly: Do you know 50K things about yourself?

On an average when you fill a questionnaire designed by one of the research agencies how many attributes do you think are considered? At best 20-25, right? And that too based on answers you feed them. You can tell them you rarely go for international vacations fearing disclosing too much about your income, but Facebook clocks a data entry every time you sign in on a new location, be it while checking in on that cafe in Paris or just scrolling through the mobile app in Brazil while waiting for your flight. And it is not just the veracity of the data, it is also the impeccable depth of it. From the kind of medication you are ordering online via their banner ads, the type of events you are more inclined to RSVP, to the genre of content headlines you are most likely to engage with. Of course, their legally murky deep dive into our personal conversations through IoT devices like Amazon Alexa and Google Home is a different article altogether. But coming back, if asked on a survey, most of us will not actually have these trigger answers ourselves. They have been fine-tuned over the years with millions of permutations and if tomorrow they decide to make virtual AI of each one of our personalities, I bet it would be eerily similar.

Relevant extrapolation

Research works on extrapolating its findings with a sacrosanct belief that the rest of the universe will also react similarly. And in 2020 where every person is laced with a hundred layers of niche sub-traits, it is difficult to expect reflecting results. You can drive the same car, have the same salary and degree from the same college, yet one of you only buys designer clothes while the other is a strict local-brand person. With normal physical research extrapolation, a lot of it is based on good luck and a lot less science. While with social media research and targeting, the accuracy sometimes can be shocking. Majorly because of the use of machine learning that adapts and evolves its sampling in real time.

Now imagine you are a luxury mall and want to bring in more shopping footfall for the Valentine’s week. Facebook and Google provide lookalike and similar audiences respectively to extrapolate from your CRM data. And you can choose it to be as specific as possible. If you have an excel of 85K loyal shoppers of which only 2K are aged between 24 and 35 who have shopped more than $10K annually, then just provide those details to either of the platforms. Facebook for example then will start showing the ad to the friends of your customers and every day it will keep filtering till it understands who exactly finally converts from that list and then only show to them and people closest to the data points of those people. It is as scientific and real-time extrapolation as any brand can dream of.

Planet sampling

Arguably the most famous marketing research done of this generation is the Dove Campaign for Real Beauty. The 2004 global study drew conclusions from 3,200 women, aged 18-64, from the U.S., Canada, Great Britain, Italy, France, Portugal, The Netherlands, Brazil, Argentina, and Japan. Dove commissioned three more studies in 2005, 2006, and again in 2008. The big, ground-breaking insight was “Only 2% of women worldwide would choose to describe their look as beautiful”.

Facebook in 2018, commissioned Accenture to find insights from their data of people who have purchased a food product in the last three months for its ‘CPG Consumer Study’. Striking revelations like: “68% of 18-34-year-olds in the U.S. agree that time can be more valuable than money.” and “40% of food shoppers in the U.S. use their phone to check their digital grocery lists when in store.” were mined. It took less than a month.

And yes, you might not have access to Facebook’s entire universe if not always generously spending on their ads nor the ability to manually gauge through all public content for insights. That is where digital listening and audience intelligence tools like BuzzSumo, Sprout Social, and Meltwater come in, starting as low as $99 a month. They will find that needle in the haystack for you and even analyse consumer sentiments in seconds with their trained software meant to scan every public data available online. Now the whole world on the Internet can be your sample.

Turnaround during trends

Researchers in the pre-social media era took years to properly analyse the global icon a children’s novel Harry Potter has truly become. Touted as an instant bestseller, it took more than a year for the book to reach just from the U.K. to the U.S. Cut to 2019, Apple TV launches with an ambitious Star Wars spin-off, The Mandalorian, and within a fortnight of launch, the whole global marketing campaign and PR seedings were streamlined with Baby Yoda at the frontlines. Why? As of July 2020, there are 600K Instagram hashtags of the show #themandalorian and well over one million for just #babyyoda. If you are a brand, you already know you need to ride the trend wave with an acute sense of urgency. And only social media research can and in the future will allow that.

Take another example of Netflix, a platform that has been bullish on trend marketing from inception and its summer hit, To all the boys I loved before. For them to make memes in the U.S., Spain to India on Peter Kavinsky, everyone’s dream boyfriend, they need to know about that insight when it is happening, not data mined by a team of research analysts after a month. They already would have moved on to Money Heist’s promotions by then.

Between a pandemic if you want to readjust your strategy, social media research again becomes the only answer for credible insight. Google just by comparing online buying behaviour between March 29 and April 4, 2020, vs. March 22-28th, 2020, could factually prove that searches for “in stock” have grown globally by 70%. Which new SEO keyword attached to your product do you think you need to pump up the ad spend ASAP?

Look at the last quarter as a test case. No one saw traditional TV ads or highway billboards outdoors, subway posters, or participated in mall activations. They did not participate in any traditional marketing research study stating things they want during the lockdown, yet the purchase funnel was taken care of completely by digital, as if on auto mode. Google knew the micro-moments customers were looking for, Facebook their purchase history, and Instagram the list of individually engaging content. All three constantly refined the researched reach in real time for better results. On a lighter parting note, if social media research can sway presidential elections across the world, as a brand manager be assured the data needed to sell smartphones or shoes could not have been passed onto more powerful hands.

Views are personal. The author is chief strategist and founder of Salt and Paper Consulting. He is also a visiting faculty of market research and campaign planning for the Advertising & PR department at Indian Institute of Mass Communication, New Delhi.