Research and Insights in 2023 – Ray’s Predictions

Globe suggesting the future

Photo by Drew Beamer on Unsplash

Ray Poynter, 6 January 2023


I spend quite a bit of my year consulting with organizations about the future and helping with a wide range of projects that try to assess current trends and future possibilities. Here are my thoughts about ‘where we are right now’ and where things might be heading in 2023 and beyond.

Where are we now?
Before I consult with people about where we are going, I always start with ‘where are we now’. Without knowing where we are, it is highly improbable that we will know where we are heading and how to handle it.

In brief, drawing on sources such as ESOMAR, MRS, IA, GreenBook and NewMR, this is where Insights and Research were at the end of 2022, going into 2023.

  • We are a large industry, over $120 billion.
  • About 50% of this is NOT traditional market research (e.g. is not surveys, is not focus groups, and is not depth interviews)
  • About 50% of all research dollars are spent in North America, with around two-thirds being spent in English-speaking countries.
  • About 50% of research projects are conducted internally by clients.
  • The fastest growing sector is ResTech, driven by the growth of platforms, especially DIY or self-serve.

Key Drivers
The key drivers are already in place and will shape our industry for the next year or two. These drivers are:

  • The global economy, especially inflation. Prices will keep changing (mostly increasing), exchange rates will be less stable, interest rates are high and could go higher, and consumers are under pressure and will in many cases change habits.
  • The growth in platforms is relentless. Many (i.e. most of them) will fail, but the sheer number of new platforms and new options will be destabilizing – and some will be successful.
  • The shortage of talent. In many fields and in most countries, there are simply not enough good and experienced people to fill all of the insight roles. This is likely to cause wage inflation, shortages, and drive the need for automation even harder.
  • Interest in AI, ML, the Metaverse, Web3, and conversational AI. Things like ChatGPT have really kickstarted the interest in the cutting edge of tech in 2023. Don’t expect too many actual products or changes – but do expect to see investment, presentations, ideas, and experiments. Move too fast and you could be the next FTX or DeLorean, move too slow and you could be the next Kodak or Nokia.
  • Data protection and privacy regulation, driven in part by data breaches, data abuse, and domestic political concerns. The European GDPR was not the end of a journey, it was the springboard in to more and more regulation. The rules are going to become more and more onerous, year by year.

The Potential Data Quality Disaster?
One item that I think needs highlighting is the risk of a Data Quality Disaster in 2023. In 2022, there were a large number of studies and reports that showed the massive problem with bad and fraudulent data in online panels. The Insights Association has been particularly active in showing the problems and risks. Studies report that the amount of suspect data that make it through to the buyers of data can be as high as 20% or 30%.

The panel companies are fighting a massive battle against fraudulent takers of online surveys, but maybe they are fighting a losing a battle? The rise of AI tools will make it ever harder for panels to screen out bogus survey takers.

This problem has been an open secret for many years, think back to reports from Grey Matter in 2012. However, the problem has proved intractable, and we seem to be increasingly sure that the problem is large enough to cast doubt on the results of research.

My big fear in 2023 is that one or more clients will have a marketing or strategic disaster and discover that that the research they used contained large amounts of bogus data. It might not matter that the direction of the research was little impacted by the bogus responses, nor that the research was the main problem causing the marketing or strategic failure. The existence of the failure, the line of evidence to the bogus data, and wish to litigate would be enough.

I think the place where the problem will happen most often is in the large self-serve, automated platforms, where non-researchers load materials for testing to a platform, press a few buttons and wait for the results. In these cases, no human is involved in selecting the sample, no human is involved in checking the data, and no human assesses whether bogus responses could have changed the results.

The fear about panels will tend to drive research towards communities, CX and non-survey alternatives.

Hot Topics for 2023
I think the five key things to watch are.

1 Even more, cheaper, and easier-to-use platforms
The growth we have seen in the number of platforms over the last few years will accelerate further this year. This will clearly power more self-serve / DIY solutions and options. It will also expand the range of things that agencies and new providers are able to offer. Clients will have opportunities to reduce costs and increase output. Agencies will be able to leverage platforms to increase their efficiencies, to help mitigate the talent shortage.

2 More tactical research, some with fancy names
Most companies only need strategic research periodically. The need for strategic insights grows in a linear sort of way. However, tactical needs happen all the time, for example, which ad to use, which UX feature to optimize, and how best to layout a retail store department this season. With the growth of platforms, the democratization of insights will grow, we are seeing a geometric growth in the use of tactical research. Tactical research is not very sexy, so we often see it extolled as agile research or as part of the build-measure-learn method. One common feature of most tactical research is that it tends to be simple, simple to apply and simple to interpret. The growth in research will mostly be in simple research, not complex research.

3 The rise of open responses, as opposed to closed responses
The bedrock of research for the last 100 years has been closed questions, for example, a Likert scale. They are consistent and amenable to simple statistical analysis. The trend now is towards open responses. We are seeing more video, more conversational AI, and more text analytics. One of the key things you should do this year is to review your use of open responses, including in your surveys.

4 Specialists and services
The growth in platforms, the democratization of insights, the shift to simpler more tactical research makes the role of specialists and those offering services even more important. The bulk of the revenue will go to automated, simple research, but the demand for specialists will remain strong. In terms of specialisms, I am talking about things like qual researchers, data analysts, business consultants, storytellers, semiotics, market modellers, etc. In terms of service, I am particularly thinking of customer success managers – the people who can help clients be really successful with the platforms they buy. This underpins the need for training and skill acquisition.

5 New faces
I think we will see a growth in the number of new faces in the global insights ecosystem, and a shift away from the more established thought leaders. There seems to be a burst of energy that reflects our emergence from the pandemic years. Of course, most of these faces have been around for a while (especially in their home countries), but they are now gaining a bigger audience. Examples of the people I am talking about include (with links to their LinkedIn profiles): Kendra Speed, Priya Lobo, Olga Keating, Alex Orlap, Hannah Rogers, Molly Strawn, Sami Wong, and Alexandra Kuzimina.

Who would you add to this list of New Faces to watch?

6 thoughts on “Research and Insights in 2023 – Ray’s Predictions

  1. Ray, thanks for sharing; that’s insightful.

    I bet that out of the tech trends listed, Metaverse and Web3 will not, in fact, attract much interest because their real value is still far, far down the line (and we can debate if Metaverse will have any). Let’s revisit this in 2024 and see if I was right 🙂

    Conversational/Generative AI will overshadow everything else when the new models, like GPT-4 from OpenAI and, potentially, LaMDA from Google, are released for public access later this year. These and other large language models will start disrupting the insights industry, especially on the tactical research side that you mention.

    I believe you’re correct regarding the prospects of the proliferation of new tools and platforms. I think it will be the most evident with new, easier-to-use point solutions and tools for specific types of research (may I mention YouScan for visual insights from consumer-published images, as an example).

    And with the looming prospect of Data Quality Disaster, clients need to make sure they cross-check their panel-derived insights with other data sources & methods (e.g., social listening) before making strategic decisions.

  2. Thanks for sharing your predictions, Ray! Also, thanks for the mention amongst these remarkable individuals, Ray! This is very humbling. Some thoughts and builds to your forecasts:

    Generative AI will steal the show in 2023! This is a new kind of AI that can spot patterns in sequential data like the words in a sentence. It is not super new, it was launched in 2020, however, the shift this year will be around accessibility and efficiency. And beyond that, we will probably see a decade-long wave of innovation built on generative AI. These incredible linguistic capabilities will have vast implications for client brands, research organisations, and us as individuals. It will definitely start to transform consumer expectations of digital interactions with brands. Generative AI will be the researcher’s secret weapon, a game-changer for supercharging existing methods.

    Metaverse – agree with @Alex Orap, hopefully, the hype around the metaverse will cool off.
    But it is important to remind ourselves that what we see now are only fragments of what will eventually become the metaverse – but some have a tendency to believe that the frag¬ments we see today are true metaverses. This leads to ’meta¬washing’ – claims about being part of the metaverse based on simple immersive or blockchain-based elements. However, virtual worlds promise to make possible experiences that you simply can’t have via a screen. Therefore, I wouldn’t bet against the metaverse DREAM itself just yet.

    Even more, cheaper, and easier-to-use platforms – In this saturation of choice of cheap platforms, it would be crucial for clients to be able to understand the benefits and limitations of these tools, sometimes it is not easy to cut through the sales spiel and understand if the tech really delivers, clients are oftentimes unable to dedicate the time and resources to proof-of-concept piloting. Therefore, partnerships with “honest brokers” will be even more important. One example of such a partnership opportunity is the NOVA Partnership initiative https://info.mmr-research.com/en-us/nova-partnership-initiative-2022 Please share if you know of any others.

    The rise of open responses delights me! I used to love coding open ends even manually in Excel back in the day. But now the rise of NLP tools and their relative affordability makes researchers’ job more productive and dare I say enjoyable! Again, some tools are better than others, some are built specifically for MR, while others are more CX.

    I agree that the risk of a data quality disaster is a concern that does not seem to be widely and openly discussed enough. If the scenario you describe, Ray, to occur, it could seriously damage the reputation and trust of the industry.

    I wonder how can we address this shared challenge? Could this shared problem create an open innovation eco-system within the insights industry where we can all share the fruits of the innovation?

    If a solution could be shared for the benefit of all, imagine, how powerful that could be across the next decade. Think Volvo sharing their three-point seat belt or more recently the dating app Bumble creating an open-source AI to stop unsolicited nude pictures.

  3. Ray, thank you for putting together this interesting prediction about 2023. I found myself constantly nodding to the points you have addressed in this article.

    As a matter of fact, the application of the qualitative method is still considered young in China, and it has not yet featured high on clients’ priority list, especially local Chinese clients. In their eyes, numbers tell a problem. But from a qualitative point of view, insights let us know why Consumers don’t need to find a voice, they already have a voice. Their side of the story needs to be heard and clients need to be encouraged to listen, even though it may sometimes involve hard facts. Therefore, when you mentioned about data quality, I couldn’t agree more. The most fundamental part of QUAL research is the quality of respondents. We need people who are honest about their opinions and not afraid of sharing it.

    Another point that I resonated with is about the more tactical research aspects. Simple research doesn’t mean simplified research, and it is important for professionals to share a comprehensive and yet, deepening study structure. It’s our job to make the interpretations that are easy to understand and apply them to our client team.

  4. Thank you for sharing your thoughts on 2023 and thank you for the honorable mention.

    You raise many good points and I agree that the automation and the self-service corner will continue to grow. As you say, many will come and many will fail, but this is just the nature of things in fast markets. What seemed to be the norm a few short years ago like fieldwork taking a month, just isn’t acceptable anymore in most cases. Therefore, solutions which allow you to reach consumers quickly, better yet cut through the unnecessary layers within the organisational structure just to get an answer to your standard question, are increasingly recognised as necessary. There are many typical questions for which you do not need to be a researcher. Marketers, product people etc can and should be able to sense check their thoughts as and when they need to leaving the more strategic work to the experts. On Fastuna the users can’t change the questions of pre designed solutions, for example. We have specifically developed them as set solutions. So this adds an extra safety net for those less familiar with hashtag#mrx . That with a complete DIY solution for those bespoke problems gives us a good balance of automation and flexibility.
    To address your point on the quality of respondents in self service research solutions. Of course you need an accountable human element. Panels are great and offer amazing research possibilities but the quality of respondents has always been a challenge even outside of automated platforms. Selecting panel partners carefully is key to ensure that real people answer questions. I know they have multiple quality checks in place to verify respondents and their answers, such as verifying IP address to avoid duplicates etc etc. We, at the same time, clean the data if necessary at the other end. We also sense check our projects manually before every launch and have quality steps in place to maintain high standards of output.
    In terms of looking ahead, there is so much information out there today. We live in the age of information and there is a war for our attention. There are still researchers who continue to write 20-30 minute questionnaires. But I feel that it is our responsibility as an industry to make the process as easy as possible for everyone, i.e. simplicity and speed where possible. Only necessary information and no nice to haves.
    Additionally, research should be used alongside other data sources. We need to continue to move towards being more integrated with other information sources and people. How to get more holistic insights by linking different sources together.
    You often talk about what the future market researcher should look like. And this ties in nicely with the more cross functional individuals to cover the gaps between specialisms. For example, we also have to become more and more comfortable with big data. What relationship do you think market researchers will have with things like data analysis and data science, for example?

  5. Thanks, Ray, for this interesting outlook into 2023 which made us very happy as it was very enlightening for several reasons:
    As experts on sensory emotion research for concepts, products and packaging we most appreciate “The rise of open responses….” , because we experienced that consumers can be motivated with very specific methodologies to express their innermost feelings, experiences (childhood-)memories. But we also perceived the limitations of such time and cost consuming methods in terms of scalability and automation. That is why we stopped offering these ad hoc services and “collated in a smart move” (Gregg Archibald) our qualitative data – collected over 20 years – into a database and transformed our research company into a digital company. The database is now the foundation of our SaaS platform to unlock qualitative data at scale. (That is why we will have a new website published soon)
    You wrote “My big fear in 2023 is that one or more clients will have a marketing or strategic disaster and discover that that the research they used contained large amounts of bogus data…”. I would like to add my personal view: since years the industry is confronted with a product failure rate of approx. 80% of CPG launches means that a lot of clients have already “a marketing or strategic disaster” not because of bogus data but because of research approaches dealing with only one part of the marketing mix i.e. testing either concept, product or pack and not the holistic context. Holistic ResTech solutions should be out there aiming at improving the chances of success in an automated and scalable way. So “Success enhancing tools” would I like to add to your hot topic 2023 list.

Leave a Reply

Your email address will not be published. Required fields are marked *