A client is going to launch a new product soon and they want to have a better understanding of their market. You're the person they turn to to get these answers. So, you decide to run a market feasibility study with an online market research firm who will do most of the legwork for you, from programming to sampling to tabulation.
Life is good. Until the MRX people start talking jargon and you’re like:
All of a sudden, the costing changes. The fieldwork is taking longer than expected. And then, you end up with a handful of issues. Could these issues have been avoided?
Well, no need to worry! Brush up on basic lingo in the world of online market research with this list.
Let's Talk Rates
Before a project begins, your panel partner will need to check feasibility. That is, whether or not they can provide all the samples you need. During this stage, you’re likely to encounter these numbers:
1. Response Rate (RR) or Participation Rate (PR)
This is the % of the panel that actively participates in surveys. This number will vary by country, by study type, and even by target audience.
Let’s say, for example, that you need 100 samples of moms with kids aged 2 years old or younger. All respondents that fit that profile can be reached, but not all of them will want to answer your survey. Some of them just won’t be in the mood, don’t fancy the topic, or unavailable due to a local holiday. The RR approximates the number of people who will want to engage with your survey.
HERE’S A TIP: When choosing between panel partners, compare their response rates. You’ll get the most out of your money in the long term when you partner up with smaller panels with highly participative members, rather than bigger panels with unengaged members.
2. Incidence Rate (IR)
The next rate to consider in calculating feasibility is the IR. This very important number indicates how many of the profiled respondents actually qualify for your survey.
For instance, there are 500 available samples of moms with kids under 2 years AND who want to participate in your survey. Among these women, the ones you really want to speak to are those who use a specific brand of baby diapers. The IR estimates the number of people who use that specific brand.
IMPORTANT: The lower the IR, the more expensive the research project will be. A low IR means that your target audience is harder to find online which means that you’ll be paying more for less samples.
So, make sure to give your MR partner all relevant information about your research goals. They'll be better equipped to estimate how the IR can behave during fieldwork.
Online research firms like us are very serious about panel management and taking care of our respondents. If panelists are not given appropriate incentives, data quality can suffer. The way we track how much incentive is given to whom is through proper redirection, via pixels or landing pages. Here are the types of end links you need to know:
When the qualified respondent answers the survey all the way through to the end, they are redirected to the Complete Page and are called “completes”.
These respondents fit your demographic target, but they don't qualify for your survey. Let's go back to the diaper example. All moms with kids under 2 years who use a specific brand of baby diapers will qualify, and those who don't use that brand will be screened out. Screenouts receive less incentives than completes.
This happens when you want to set up quotas. For example, you’d like to set quotas for age and want to collect a 40:60 ratio of completes for moms in their 20’s and 30's. Once 40 completes for the moms-in their-20's quota has been achieved, additional moms in that age group who try to answer the survey will be tagged as Quotafulls.
Once data collection is wrapped up and all the responses are in, it’s time to check data quality. You can do this yourself, or you can ask your MR partner to do it for you. Generally, these are the types of bad samples you should watch for:
6. IP Duplicates
In some cases, you'll find that two or more respondents share the same Internet Protocol Address (IP Address). This can happen when the same device or the same computer network is used to access the survey more than once. Exclude the duplicates from your data. There's no way to tell if they're really unique responses from different people, or the same person trying to answer the survey multiple times.
When the Length of the Interview (LOI) of the survey is 30 minutes and the respondent answers it in 5 minutes, they’re what we call “speeders”. They raced through the survey, probably without even thinking about the questions. In this case, best to remove them from your data as well.
RULE OF THUMB: If the respondent completes the survey in 30% or less of the estimated LOI, they're potentially a speeder.
Then, there are respondents who take an hour or more to answer a simple 30-minute survey. These are the “outliers” and you should definitely take them off your data. For all you know, they might have been researching about the survey topic while answering!
There you have it. These are 8 basic market research terms you should be familiar with before you take on a research project online. The next time you hear these terms, you'll be ready.