Dynata: One Company. More Data. More insights. More results.

A Market Research State of the Nation Review

Significant trends around the global economy, automation, AI, chatbots, research pricing, and client demands will affect MR in 2019.

Editor’s Note: We have featured several articles on the latest GRIT findings and implications in the past few weeks.  Today, we have Ray Poynter’s POV, placing these findings into somewhat wider context.  As usual, Ray has a number of wise things to say.  I am particularly struck by his comments on how a slowing global economy may affect market research in 2019 (to say nothing of Brexit), and by his client-side insights.  Great stuff!


At this time of year I am busy running workshops with both agencies and client-side organizations where we evaluate where the insights profession is at the moment and where we are heading. This post summarizes the key themes that I have been sharing with my customers. These themes are supported by such invaluable sources as the GRIT Report, ESOMAR’s Global Market Report and Global Pricing Study, my own work with NewMR and of course my involvement with IIeX (the next IIeX being in Amsterdam February 18 & 19).

It’s the Economy Stupid

At this stage we can’t be sure how the global economy will perform in 2019, but there is clearly a major risk of a bad year. The economy is the key driver for 2019, if things are OK, we will see some cautious changes and some growth in some places. If 2019 is bad, then some clients will go bust (in the retail sector many already have), many clients will need to dramatically reduce their marketing and market research budgets, and we will see more supply-side failures and distress sales/mergers.

Many of the researchers and clients I speak to are optimistic about 2019, and the GRIT Report shows that the majority of participants in the study are optimistic too. However, I think they are misguided. If I ask a pair of questions, ‘How well will you do next year?’ and ‘How will others do next year?’ there is a big mismatch – people recognize that others will struggle, but they feel they can beat the house – my feeling is that the house generally wins.

Commoditization of Research

The ESOMAR Global Prices Study shows that the price of online research projects that do not include non-standard elements is becoming cheaper and cheaper, a process that has been going on for ten years, and which shows no signs of slowing. This trend is a result of automation, better product design, and a hyper-competitive market. Research is not different from cars, hotels, postal services, banking etc, anything that can be commoditized will be, making it cheaper, faster, and less error prone. The process also creates a market for non-standardized solutions for things that can’t be commoditized.

Automation and AI

2018 was a big year for automation and AI, and 2019 will be an even bigger year for these two. Automation has been a key trend for several years and AI is driving the next stage of automation. If the economy is bad, the key use of AI will be to create cheaper/faster options. If the economy is less bad then AI will also be used to create more innovative options, for example in the field of video. In the GRIT Report, Stephen Phillips (CEO of Zappi) and Zach Simmons (CEO of Discuss.io) make the important point that the key to successful automation and AI is scalability – the winners in 2019 and 2020 will be those who prioritize and deliver on scalability.

If you look at the techniques that are lower down the GRIT rankings for Emerging Technologies (things like VR, neuro, BE, and gamification) you will note that they have not leveraged automation to create scalability. The technologies at the top of the list, online communities, mobile first surveys, social media, big data, text analytics are all heavy users of automation to achieve scalability and are increasingly using AI to enhance that scalability still further.

One of the challenges for the use of AI in market research is the tendency to dismiss any technology that works as being ‘not really AI’. For example, before we could do automated sentiment analysis it was considered to be AI, but now the technology is largely working it is often dismissed as not being AI. This approach to AI disguises how much AI is already being used, and this tendency holds back the enthusiasm to use more and more AI to automate more and more things.

And Chatbots

One particular application of AI that I think will see a boom in 2019 is the use of Chatbots. These are already in widespread use, with a large and growing number of providers. One problem I have noted is the tendency for some people to say ‘Hey, that is not perfect, we should not use it.’ – these were the people who also said we should not use online surveys, and who happily field old-fashioned surveys that are anything but perfect.

The Campaign for Real Qual

It is clear that the growth in big data, programmatic solutions, and passive data collection are leading to a growth in the demand and utilization of qualitative research. One interesting contrast is the difference that the industry audit from ESOMAR estimates that the Qual/Quant split is about 15%/85%, but the GRIT study (which focuses on the more innovative, faster-moving parts of the industry has a ration of 30%/70%. At the moment, the majority of this qual is still traditional (e.g. focus groups and depth interviews) but the demand for larger-scale, digital options seems immense (providing it can be cheap and fast).

There is also good business to be had in very high quality qualitative research. This is not scalable, is only used when the business problem is important, can’t be solved cheaply, and where there is time to solve it properly. This is never going to be a big proportion of the research pie, but the margins are good and the work satisfaction of artisans and buyers of artisanal services tends to be high. Expect this sector to grow and profit, but not to be the dominant sector. Consultancy falls into a similar small but sweet spot. Most research is tactical, but that is becoming a commodity, consultancy is required less often, but the margins and satisfaction tend to be higher.

Client-side Insights

People like Stan Sthanunathan have been talking about the need for change for several years. Over the last year or two we have been seeing more and more signs of the changes. One illustration from the GRIT study is the mismatch between what clients are buying and what agencies are selling – in terms of Big Data Analytics and Social Media Analytics. This gap indicates that clients are buying large chunks of their research from companies that are not ‘market research’ agencies.

The key drivers for most companies have been cost, but speed is becoming an even bigger driver. Stan Sthanunathan has for a couple of years been spelling out the message by saying we have to “double the impact at half the cost”. When I interview clients and look at their path to purchase what is clear is:

  1. Clients keep saying that 80% right is good enough.
  2. If they like the solution, they will pay 20% more for it (that is 20% more than the lowest price, not 20% more than the highest price).
  3. But, if the solution is late, it is 100% useless, and it makes the insight professional look bad.
Client Concerns

The GRIT Report (and my own interviews) show that clients are relatively satisfied with the way that the mechanics of research are conducted – for example survey design, fieldwork, and analysis. The biggest shortfall is the relationship between the value of the research and its cost. The main issue is not the cost, it is the value relationship between the cost and what is delivered. We get an indicator of what clients mean by this from two other factors that they are unhappy with, the quality (or lack of it) of business insights/recommendation and data visualization. However, if the economy in 2019 sucks, it is hard to imagine this improving much – which will further strain the relationship between clients and agencies, and encourage even more clients to look to non-MR sources.

My Recommendations?

One risk when making predictions is that people can often include things that think ‘should’ happen, along with the things they objectively think are happening. So, here are my recommendations, I do not think that everybody will deliver them, but I think they should try:

  1. Plan for growth, but be prepared for setbacks. Your clients may pay even later, some of your suppliers may merge or go out of business, and clients may scale their plans back.
  2. Prioritize scalable automation, focus on making things easier to use, faster, and cheaper. If you are a software platform, try to make the less able users of your software more likely to keep their job, because you enable them to produce more/better outputs. If you are a client-side insight function, produce outputs that do not need to be explained, which are available quickly, and which do not eat up budget.
  3. Embrace AI, and call it AI. AI is going to change our industry (and every other industry) so be on the winning team.
  4. Employ Client Success Marketing, by which I mean making the individuals you deal with more successful. By all means make P&G, or Unilever, or Coca-Cola more successful, but do it by making the people you work with more successful. Find out what each person’s value proposition is and deliver to that.
  5. Nobody has a clear handle on what good data visualization is. You can’t buy it or take a course that will teach you it (both will help though). Ask your client for examples of data visualizations that work for them and build around those.

Remember, if you or your company would like a bespoke version of this material; just contact me, we have options for face-to-face workshops and online versions.

Please share...

3 responses to “A Market Research State of the Nation Review

  1. Thanks for the useful commentary Ray. In particular, I like your perspective on building scalable solutions and and leveraging AI to do so. Another thing you point out is that non standardized work is perhaps immune to commodization. I believe one opportunity is to drive solutions that are more configurable through highly flexible software and services so that some of this non standard work can become more scalable, and enablers like AI can be leveraged. So perhaps not commoditizing it, but addressing the speed and cost concerns clients have. Thanks again for a great post.

Join the conversation