Research Note #2: Is COVID-19 impacting Market Research Supply and Demand?


COVID19
Thought Leadership

Cint Platform Data Shows No Significant Impact (Yet) of Coronavirus on Industry Supply and Demand

As the largest of the industry’s sample exchanges, we have the enviable position of being able to witness the ebb and flow of market research supply and demand. We are sharing this data to provide greater understanding of the impact of the pandemic for decision-makers on both sides of the market research “transaction.”

For our thoughts on the impact of COVID-19 on consumer behaviour, please see our Research Note 1: Monitoring Changes in Consumer Behavior Due to COVID-19

 

Background

The rise of the COVID-19 and the subsequent “lockdown” of whole populations is expected to negatively impact the global economy. The big unknown is how long this will last. Decreased movement is driving down consumer spending. Companies are starting to retrench by scaling back expenses. Now, everyone is asking questions that revolve around the impact of the virus on supply and demand in market research.

 

Demand and Supply: The Market Research “Transaction”

To borrow from the terminology of economics, the market research transaction has two parts. Demand in our industry is reflected in a brand or company’s desire to learn more about consumer behavior. We can measure demand as a simple function of the number of consumers from whom these firms wish to collect behavior or opinions. We can express this as we might a consumer packaged good, by looking at the number of projects run and the number of completes per project.

Supply is measured by the number of respondents available, though in today’s world the concept of invitations and response rates is increasingly archaic. At Cint, we do have invite-driven suppliers on our platform, but we have more supply that is programmatically driven. For the latter, we can look at entries into the platform. (An entry is effectively a programmatic respondent being sent from a supplier and touching our platform.) Ultimately, demand and supply meet at the point of the interview and subsequent complete.

 

Hypotheses

To study the impact of the virus on our industry, we are looking at both demand and supply factors. As researchers, it helps us to state hypotheses to ensure that (a) we have a comprehensive view on the situation and (b) we have clear indicators by which we can gauge impact.

 

Demand hypotheses:

  1. Clients reduce MR spending, resulting in fewer projects commissioned.
  2. Clients reduce MR spending, resulting in fewer projects and fewer completes per project.


Supply hypotheses:

  1. Respondents become less/more likely to participate in research studies, causing their numbers to decrease/increase, as measured by either:
  • a decline/increase in the number of programmatic entries to the platform;
  • fewer/more respondents to email invitations; or
  • lower response rates (respondents divided by invitations).


Joint Demand-Supply hypothesis:

  1. For whatever reason, conversion rates decline. (Conversion rate is defined as the number of completes divided by entries, which takes into account all factors that stand between a respondent and her/his desired outcome of completing a study).
  2. For whatever reason, dropout rates increase. (Dropout rates are the number of people abandoning a survey divided by entries. Dropout is typically a function of a terrible experience that is survey-related.)

 

While the demand hypotheses are one-tailed (e.g., we do not expect demand to unexpectedly grow), the supply hypotheses are two-tailed. Put differently, there are compelling reasons that could explain either growth or decline in participation. Of the supply hypotheses, the first is absolute in nature, namely that that total levels of respondents or entries will grow/shrink. The second is relative in nature, namely that the conversion rate (completes divided by entries) or response rate (respondents/starts divided by invites) will change.

The joint demand-supply hypotheses are listed separately because their causality is unclear. We are focused on overall conversion rate because this is our key platform health indicator. If the type of work changes (e.g., targeting more niche audiences), conversion could go down. If clients are doing more “nat rep” work seeking to understand broad trends, it could go up. We think drop-out rate is a good statistic to look at as well: it could go up or down based on one’s perspective about people’s tolerance for participating in a given study.

It is worth noting as well that some of these indicators are imperfect. For instance, those suppliers who have fully leveraged the features of our API will send people into the platform when there is available inventory. Thus, if inventory (studies) were to decline, we would expect entries to decline. Likewise, a peculiarity of this dataset lies in the fact that we archive data on invites after 30 days when a project is invoiced.

Finally, for the avoidance of doubt, we are not doing significance testing here. Nobody needs us to calculate 95% confidence levels for this. We are using an equally valuable and less computationally intensive method called “eyeballing.”

 

Results

We have selected a handful of countries for this analysis. Charts are available in PDF format showing various metrics graphed alongside daily figures of confirmed COVID-19 cases. In every chart, we have deleted the vertical axis labels for confidentiality. In some charts, we have also re-scaled the data. In every chart, the COVID-19 cases are reflected on the secondary (right-hand) vertical axis.

An example of these charts is shown below. These are programmatic entries and respondents from the United States.

 

Programmatic Entries and Respondents from invites - Cint

 

Below are the results through March 30:

  • FR/US/JP: No apparent impact so far on any measures.
  • ES/IT: Email respondents have declined somewhat, but programmatic entries have risen. Overall no impact on completes. No impact on other measures.
  • UK: Could be trending lower. Entries and respondents showing a slight drop. Completes were down slightly but this could be levelling off. Completes per offer and invites sent are trending down slightly.
  • DE: Respondents, completes, and projects are starting to drift downward.
  • AU: Increase in number of projects, though a bit of a decrease in completes per project. All other metrics stable.
  • KR/CN: No apparent negative impact, which is interesting because these countries are on the downside of the contagion curve.
  • IN: No apparent effects. While response rate is declining, both the number of invites and respondents are growing. The former is growing faster than the latter.

 

COVID-19 impact tracking

 

 

Conclusion

We see little evidence thus far of any negative impact on supply or demand through Q1 in all countries, though it does appear that the UK and Germany may be trending downward. In our own business, we have had a very strong first quarter which was both in line with budget and a significant increase on trading one year ago. We are nevertheless starting to see some cancellations coming in for April and May. We will be sharing platform data through the crisis, likely at the beginning of each week.


Written by: JD Deitch (COO at Cint)


Read our other Research Notes:

Liked this post? Spread the word on:

More from our blog