Seeking to do things quicker and more profitably, cutting corners without compromising quality, doing a job well enough. For many of us, these are often necessity in the modern world, regardless of the industry or circumstance we’re in. The world is getting faster. It will continue to get faster. Market Research is no different.

But this is not a new trend, nor are our attempts at leveraging technology to meet the challenge through some degree of automation. For years the industry has waged cold war against the automation of machines, believing itself an intrinsically human business and scoffing at the errors made by our computer friends. And often rightly so. Quality is key in our game. If we haven’t moved a client forward, then what really was the point?
So in a value driven industry, quick and dirty often helps no one, but we rarely have the luxury of doing long and perfectly clean either. Left to run wild, the research process can indeed be a long and arduous road. And whilst that long road may deliver an output we can be very proud of, the opportunity to deliver that all important value may well have gone, or we no longer have the space and time to deliver it effectively.

As research agencies, perhaps we’re guilty of being a little too proud of our carefully honed data collection and visualisation techniques. After all, it’s the part of the process that routinely damages our profit margin. It’s fiddly, error strewn and prone to abuse from that demanding client who wants everything cut by everything else. All too often, the energy and creativity sunk into project design (not to mention the profit) has been drowned out by the drudgery of the rest of the process, such that by the time it comes to the all-important end game of actually moving a client’s business forward, we’re all a bit worn out. And the next client is waiting. How many business leaders have decreed that their star player must ‘back off’ from a project because there simply isn’t the budget for them to be involved anymore? Our faultless execution of the process has, in the end, resulted in a ‘dirty’ output anyway.

But without a good process, there would be no good insight, so it’s perhaps understandable that insight teams have often been reluctant to look to machines. However, as automated capability evolves, the benefits of doing it and doing it right are hard to ignore. Technology led businesses are giving focus to the component parts of the research process that offer the least opportunity for profit and require the smallest sprinkling of human fairy dust, and simply taking them out of the equation. Saving fortunes in money and time.

Automation is ready to take these tasks off our hands if we’re careful with how we build our software. Sampling, data collection, tabulation and basic reporting can all now be easily handled by a machine; leaving the humans to think about smart research design and high end consultancy – the machine’s input and output – where that fairy dust is still the difference between client retention and client inertia. And where the budget sheet smiles sweetly back at us. Not only that, weeks of manpower are replaced by hours of machine processing meaning we finally deliver for clients when they need it. Agencies will evolve into this new shape. Why wouldn’t they? Their old road is rapidly aging, as Bob Dylan once said.

If automation means we can have quick without dirty; or more than that – if it means quick can actually help clear our desks and make our output even cleaner than it was before, then the differentiator becomes the degree of quality an agency can provide. The focus will shift in a way that demonstrably benefits the end user. And not before time!

Skill-sets across the research agency will shift accordingly. The ties between research and academia are likely to grow closer, as we seek ever improved methods of extracting the truth for our project designs. This is no bad thing and often the reason graduates enter the trade in the first place. Consultants that can engage a client audience, translate the insights and help them build action out into their businesses will be at the centre of every successful agency business. The industry already realises that, but the consultants need to get even better. If they don’t, client-side insight teams will find the temptation to bypass them for an automated solution, that they interpret and disseminate themselves, all too enticing.

Because client-side Marketing and Insight leads will be looking to their own balance sheet and driving their teams to skill-up and dispense with that highly paid research consultant. With quick and quality comes affordability, as machines cut out some of the time expense. Although that’s inevitably a sensitive subject, it should encourage more research done, more often, by more businesses.

Our savvy clients will begin to understand the power in testing early and testing often. Building anything on a hunch will begin to seem like outdated folly, if a quick test costs a fraction of the old world price and gives results within the day. Clients might test an ad all the way through the phases of its design. SMEs will suddenly find price points for quality research output that they can work with. At ZappiStore, we are working closely with the vanguard of this movement.

But how long will these battle lines between machine and man stay put?

Machine learning breaks new ground at a pace that the research industry is unaccustomed to. Its forays into design and consultancy can be limited and clumsy and the quality questionable, but the output gets better all the time. With news articles being written effortlessly by machines, some of which can be programmed to write with particular emotions like sarcasm, how long before they can do the job of the quality consultant? Probably a very long way. But the challenge has been laid down. Quick and quality is here and the research industry must think pragmatically about what level of quality is really required and whether a machine can lighten the load. It almost certainly can.

Jamie Shacklock

Posted by

Jamie Shacklock

Loading Disqus Comments ...