Automation Everywhere – Even Market Research
In a world where we now have driver-less cars; smart thermostats that not only program themselves, but also pay for themselves; and the ability to deposit a check just by snapping a photo it is clear automation is everywhere. But what about market research automation?
Over the years we have received many questions about market research automation. For example, we’ve been asked why images can’t all re-size themselves to a single standard size, or why open-end responses can’t code themselves automatically without having to define categories and keywords. These are not insurmountable challenges, correct?
The market research industry is increasingly taking on such challenges and automating everything we can think of. Sentiment analysis is a great example of market research automation. Sentiment analysis – which is based on text or speech analysis – gathers qualitative information about a respondent and is widely used to analyze social media in order to understand whether someone was happy, unhappy or, as Sheldon Cooper would say, “You’re being sarcastic, right?”
Here at Kinesis, we’re doing our part to participate in market research automation as well. We plan on updating our blog post from time to time with bits and pieces on market research automation, but for now we’d like to turn our focus to online panels, how they are being automated, and what some of the terminology means.
In the world of online panels, the primary areas of automation are:
– Routing / sample optimization
– Automated sampling
Routing and sample optimization can refer to the same thing. While routing has a somewhat negative connotation, when viewed as sample optimization it sounds more positive. The truth is, depending upon how it is implemented; routing can either be a positive experience for respondents or a negative one.
For more on the pros and cons of routing, please see our post on: “The Evolution of Online Sample.”
The purpose of routing, as it pertains to optimization, is to be able to use sample more efficiently. If a respondent does not qualify for an initial study, they are routed into another one for which they may qualify. In order for routing to work efficiently, respondents must be invited along with their qualifying criteria (age, gender, and other demographics), so they are matched to a project for which they will qualify, rather than being invited to projects randomly, terminated from one, and then bounced to yet another project. Optimization, when well-executed, provides the respondent a better chance of having a project to take.
Automated sampling generally refers to tools that provide end users with the ability to pull the sample themselves. They are able to obtain bids, order sample, and direct it to their project as needed. There is generally some recognition of cost savings for end users over full-price models that rely on a project manager. The majority of automated sampling tools are vendor-provided solutions. Complex sampling needs may not be suitable for automated sampling.
Automated sampling can be taken even further – to pull sample from sources automatically as needed (although this process is behind the scenes and is a part of the sharing of panelists across vendors that most end users do not see. It is also a part of the sample exchange – such as that offered by Fulcrum. The exchange is a multi-vendor approach that allows end users to compare prices across vendors and source from multiple vendors at the same time.
Finally, we have all heard of content aggregation and the concept in market research for sample is not that different. Aggregation refers to vendors that source from multiple online panels and essentially act as re-sellers. Aggregation tools typically automate the process of requesting and sourcing sample when a given panel has insufficient panelists to fill a job.
So maybe we don’t have surveys that program themselves (yet), but there are some exciting innovations in market research automation. If we had to guess, we’d say that as the world around us continues to become more automated so too will market research tools and techniques.