Why Conventional Research is Still Relevant in Times of AI

Spending on AI and cognitive systems is pegged to soar over US$ 55 billion by 2021, signaling the growing penetration of these technologies across various industries. You would be forgiven to think that the case would be no different with the market research industry, with quantitative market data giving way to deep customer insights. This, however, is miles away from reality. While there’s no denying that automation through big data analytics has certainly made rapid inroads in market research, it hasn’t quite led to a disruption of traditional research methods, and with good reason too. In fact, the human element that AI tries to replace is the key differentiator here, as at the end of the day, all goods and services are, targeted at humans, either directly or indirectly. 

AI-based Market Research Vs. Conventional Market Research

Let’s get a better understanding of how conventional market research still holds its own, as AI-based market research fails in an actual recorded market scenario.

Scenario: A region recorded astronomical rise in sales of automotive batteries in the recent past. Based on this information, a car manufacturer would like to explore avenues of growth in the region and seeks market research data to help with decision making.

AI-based Research Findings: Sales of automotive batteries in a particular region reported an astronomical rise in the recent past. AI integrated market research, in its current nascent stage, which relies on large volumes of data to deduce patterns and arrive at conclusions, would attribute this to high sales of automobiles in the region. A steady increase in sales of automotive batteries may be indicative of increasing volume of vehicles plying on the roads in the region.

Conclusion: This can be interpreted as an ideal scenario for a new car manufacturer to venture into the market, especially dealers of used cars, given the high demand for second-hand automotive batteries.

Conventional Market Research Findings: Conventional research, would factor in other aspects, which may not be evident at the surface level, such as prevalent economic, regulatory, social, and even climatic conditions. Secondary research revealed that electrification in the region was just over 50%, mainly limited to urban clusters. The region in question had a disproportionately low vehicle parc, something that would probably not be factored in during AI-based market research. Moreover, poor economic condition and abysmal road infrastructure, coupled with harsh environmental conditions in the region would be deleterious for the automotive industry. Primary research revealed that sales of used car batteries were high, with the local population relying on these batteries--that they would periodically charge --to power their homes during daily power outages that last for as long as 8 hours.

Conclusion: While the region would be highly lucrative for manufacturers of automotive batteries, more specifically providers of repurposed automotive batteries, a car manufacturer would have very little to gain in such a hostile business environment.

Very clearly then, conventional market research provides a higher level of fluidity, which enables it to circumnavigate situations where standard statistical research approaches may fail. Also, as AI is highly, if not completely, reliant on big data to understand historical patterns and make future projections, it is found to be greatly lacking when making projections about new emerging markets, niche application areas, and also when assessing a competitive landscape. Spotting and analyzing hidden data is key to accurate macro- and micro-level understanding of the market, something that conventional market research banks upon to not just deduce market numbers, but also provide invaluable workable insights that shape the course of regional, as well as global economies.

The Way Forward

Even the most ardent proponents of conventional research will admit that the approach isn’t without its fair share of shortcomings, mainly when it comes to extrapolation of market numbers based on relatively small sample groups. Moreover, exhaustive market research involves major investment in terms of time and resources employed in gathering market data and then processing it. The biggest advantage of AI is in the almost unceasing carrying out of repetitive and menial tasks. Market research automation tools simplify the process of gathering and processing large volumes of data and reaching out to a larger and more diverse sample group. Menial tasks such as finding sample, survey routing, data cleaning, and raw data analysis can be managed using automation tools. Researchers can focus on putting their rich experience and expertise, combined with a more 'human' understanding of factors that may not be directly linked with a particular market--yet have a key impact on it--on providing insights by interpreting numbers and trends identified by automation tools. Moreover, automation also helps drive down human capital costs, increasing profits and also, making market research more accessible to smaller companies. This approach of pairing, rather than pitting, man and machine is already paying rich dividends for market research players, and in turn for their clients, as evidenced by exponential improvement in business decision making based on incisive market insights.