21st Century Market Research: Process

The Assembly Line

I’ve often compared the traditional market research process to a 20th century “assembly line” approach.  

By this I mean that there tends to be a person responsible for creating a study design, which is then given to a Project Manager to be moved through a series of steps that involve handoffs to various technical experts.  These experts use specific tools (that the Project Manager doesn't know how to use) to add value by transforming the inputs they receive into new outputs.  

Each new output is then passed back to the Project Manager to be quality checked before being handed off to the next person in the chain.  This process happens iteratively until the project is done.  

For example, here is the traditional process for a typical quantitative survey:

• Senior Researcher creates a study design which includes a survey outline/draft, analytical plan and reporting plan

• Project Manager creates a final survey, gets it approved by the Senior Researcher, and passes it off to a survey programmer to get it programmed

• Project Manager then QC’s the survey program before starting data collection

• Project Manager oversees the fielding of the survey and keeps the Senior Researcher informed of progress and any issues that need discussion 

• The resulting data is checked by the Project Manager before being called final

• Final data is then passed off to a Tabber and a Coder to be prepped for analysis

• Tabs and coding are QC’ed by the Project Manager

• Project Manager uses the tabs and coded opens to create a draft of the report

• Senior Researcher reviews the report, provides feedback and analytical commentary, and delivers to the client

This process made sense given the limitations that were present historically.  For example, the average researcher trained in design and analysis did not have the technical knowledge to use data collection and data tabulation software or the time to code open ends, not to mention the expertise that comes with years of experience performing these tasks. 

The Rise of DIY Tools

The rise of DIY software in recent decades changed this process.  It encoded many of the more technical aspects of tasks such as survey programming and data tabulation into software that allowed researchers (like me!) who lacked technical expertise to handle these tasks on their own, at least for projects of low to medium complexity.

However, DIY tools have not eliminated steps, they have just made the existing process more efficient through automation and reducing the number of people involved.  

Fewer people means fewer hand-offs, which increases speed-to-insight and gives fewer opportunities for misunderstandings between team members.  

The end result is that it has been possible for some years to have a project “team” that consists of just a couple of people:

• A Senior Researcher who builds and maintains client relationships, translates business objectives into research designs and puts the final stamp on analysis and delivery

• A Project Manager who oversees the execution of the design and resulting plan by using DIY tools to program the survey, organize the data on the back end and create the report  

From DIY Tools to Co-Pilots

The wave of AI development that is sweeping through the market research industry right now is an intensification of the DIY trend.  This is leading to a future that is, to some extent, already here in the form of AI Co-pilots currently being built into DIY tools.  

But co-pilots don’t fundamentally change the process.

Project Managers are still dealing with handoffs, but they are handing off to apps instead of people as they bring data from one tool to another to complete a project. This process may be more cost and time efficient but the user still needs to be familiar with each specific tool being used in the process.

From Co-Pilots to Concierges

The next step in this evolution is likely to be apps that sit above individual software platforms and have the ability to orchestrate their activities.  Applied to the “typical quantitative survey” I referenced above, this would lead to a process that has the same structure as the traditional process, but which functions in a radically different way (similar to the way online data collection was both similar and radically different from telephone, mail and intercept-based data collection).

This leads to a process that looks more like this:

• Senior Researcher creates a study design which includes a survey outline/draft, analytical plan and reporting plan

• Project Manager uses a Concierge app to:

  • Create, program and QC the survey

  • Field, QC and finalize the data

  • Analyze the data and create a draft report

• Senior Researcher reviews the report, provides feedback and analytical commentary, and delivers to the client
  

It's not hard to imagine scenarios from there, again starting with simpler use cases as with DIY data collection tools, where the Senior Researcher or even the business owner can use a Concierge app to run a study by themselves from design through analysis.

I don't think this will happen overnight, but I do think it will happen. Timing is, as always, TBD.

Implications

To quote Professor Andrew McAfee of MIT, "When a new technology emerges, we tend to overestimate its short-term influence and underestimate its long-term influence. We're doing the same thing with AI right now." (from an interview in Der Spiegel, March 4, 2023, p. 14- my translation)

So, what can we do as researchers to prepare ourselves for a future we can't predict with certainty? Two things stand out to me:

• Ground ourselves by thinking about "how these things usually go" with the impact that the Internet and DIY tools had on the research industry being our guide

• Considering the ways in which our roles, whether on the "vendor" or "client" side, are likely to change and ensuring that we have the skills to succeed in the new era