Blockchain for Surveys

Towards a Fraud Detection Clearinghouse

Five years ago, you couldn’t swing a dead cat (sorry, PETA, we need an animal-friendly alternative for this phrase too) without smacking an article about blockchain, including in the market research world. Web 3.0 was upon us, with blockchain providing a core protocol.   

Web 2.0 is largely considered to be the “social web”—marked not just by the advent of social media but the inclusion of interactive elements on a wide variety of websites. With Web 2.0, we went from being information consumers to content creators and masters of our own multimedia universes.

Web 3.0 is the “decentralized web” or, put in a slightly different lens, the “value web.” Decentralized means that the information is hosted on individual users’ devices rather than some central server. If Web 3.0 is to come about, blockchain technology, which proliferates information to all interested parties through a decentralized “ledger,” will likely play a major role. The adjacent entities, cryptocurrency and smart contracts, will also very much be along for the ride.

By “value web,” I mean that blockchain provides the technology to facilitate trusted exchanges of digital assets, which could represent anything of value in the real world. Keywords include ownership, exchange, and trust. Again, cryptocurrency provides a well-known example, as do non-fungible tokens, or NFTs.

While it is easy to poke fun at the hype that has surrounded blockchain (and crypto and NFTs), it is an interesting and potentially useful technology, even if it doesn’t revolutionize the internet. There are certainly some interesting theoretical applications within market research.

This should not surprise anyone. Blockchain is all about trust, the trusted exchange of information about a digital asset through an immutable ledger. Market research is about truth, which presupposes trust in those providing, collecting, and distributing information.

Through blockchain technology, everyday people can share data about their interests and behaviors in an anonymous way that cannot be tampered with by others and that redounds fair (agreed-upon) compensation directly to them.

Imagine the following scenario. I take a ten-minute survey about my grocery shopping habits. With a traditional survey, I give my data (survey answers) to the client sponsoring the survey and (likely) never hear from them again. With a blockchain-based survey, I answer a ten-minute survey about grocery shopping for free, but I own the data (with a private blockchain network, data can be hidden from everyone who has not paid for it). Grocery Store A can ask to see my data. I show it to them for five dollars. But Grocery Store B can also ask to see my data, and I can show it to them for five dollars as well. In theory, I could sell my data to one hundred grocery stores and only had to take the survey once.

While this sounds fantastic for survey takers, there are a few issues. For one thing, survey questions are sometimes idiosyncratic, meaning they elicit information that only the survey developer/sponsoring client cares about. If I take a survey about windshield wipers, there aren’t going to be many companies banging down my door for my data. Just other windshield wiper companies. The catch is that if I am Windshield Wiper Company A and I want to run a survey of US car drivers, I would probably rather deploy a survey in the conventional way than on a blockchain because I don’t want other windshield wiper companies to easily get the same data and see what I asked on the survey.

Another issue has to do with timeliness of the data. Survey data has a rapidly diminishing shelf life. Windshield Wiper Company B might come across survey data for sale on the blockchain that was commissioned by Company A, but if the survey was three years ago, they probably wouldn’t want to buy it. As someone taking surveys and building up data to sell, my supply is constantly going bad, and there’s a question of whether it’s worth it to maintain data resources like this.

Lastly, there is the tricky issue of anonymity. With cryptocurrency, one of the key features is that anyone can create an anonymous account. From the perspective of survey takers, anonymity is great. You could provide sensitive information about yourself without anyone ever being able to trace it back to you. One of the cool things about blockchain is that you can also build a reputation as a reliable and truthful survey participant, and any survey company can verify this before sending you a survey.

The catch, as with all online survey panels currently, is that you could say you are a US consumer but be in China using a VPN. Furthermore, you could create as many accounts as you want and make money by taking surveys all day. This is undesirable from a data quality perspective. Again though, this is not a problem that would be particular to blockchain.

As you can see, there are a few reasons why blockchain surveys don’t really exist. By blockchain surveys, I mean surveys where your data (survey answers) are encoded in the blockchain. If you decide, after reading this post, to start a blockchain survey company and end up making millions, please say nice things about me.

With that said, there are blockchain-based market research companies. Three that I know of are Measure Protocol, Opinion Economy, and Veriglif. The concept is the same as with blockchain surveys—you own your data and sell it to brands through a blockchain protocol as you see fit. It’s just that the data in consideration goes well beyond survey data. With Measure Protocol, for example, you are essentially selling data about your online behaviors including your Google searches and your visits to certain websites. This data is more clearly useful and is always fresh because the tracking feature of the app you use is always turned on.

I should note that there are similar activity-tracking apps, such as Survey Junkie, that did not base their technology on blockchain, and you might wonder whether that matters.

Consider blockchain developer Gideon Greenspan’s rules of thumb for qualifying a blockchain use case. According to Greenspan, the data at issue needs to be updated by multiple parties that cannot be assumed to trust each other. Furthermore, all transactions should be dependent on prior transactions within the system. Importantly, use of the technology needs to be more efficient than an alternative approach.

If you want to share your smartphone activity through an app in exchange for a financial incentive, I’m not sure there is a need for blockchain technology to be involved. The data is flowing in just one direction—from your phone to the app company and then to their clients. I would also guess that blockchain would not be a cheaper base on which to build the app.

I wish I could provide you some sense of how successful blockchain market research companies have been. Unfortunately, there is not much news available and certainly nothing related to company revenues or market value. I can only say that I had not heard of these companies prior to researching this post. Take that with a grain of salt.

Now you may be wondering, have I led you all this way just to discount the idea of blockchain as a key technology in market research? In fact, I have one use case that I think might work well, and it is related to online surveys.

There are within the market research world two relatively well-known “fraud detection” services for online surveys. There is Research Defender, which was acquired by Rep Data in October 2023, and there is Imperium, which is owned by the behemoth survey panel company Dynata. These companies essentially track the survey-taking activities of users across all of the panels with which they partner and can flag bad actors. The users are tagged or identified by their digital fingerprint, a technique that allows you to know that, for example, a member of Panel A is actually the same person (or very likely using the same device) as member X of Panel B. If Panel A finds this member to be a bad respondent, the fraud detection service can alert Panel B as well, allowing them to remove the same person from future studies. You can also use such a service to, for example, say that you don’t want to invite anyone to your survey who has already attempted five surveys that day. Again, the value is that this information spans multiple panels.

One of the biggest weaknesses of these services is that their knowledge of panelists’ activities is limited to the panels with which they partner. When it comes to the universe of essential survey fraud data, Imperium has part of it, as does Research Defender. You can pay for both services, as some companies do, but the data is never completely bridged. You need separate quality control rules for the separate services.

Here is my radical idea: the market research industry should create a non-profit clearinghouse for all of this data so all online survey activity, no matter the panel, gets tracked and is available to all research companies participating in this clearinghouse.

This approach has worked well in the education sector, where students’ enrollments and credentials are tracked across nearly all higher ed institutions by the National Student Clearinghouse. If a student has applied to a grad school and claimed to have a bachelor’s degree from University X, the grad school admissions staff can verify this quickly through the Clearinghouse. Alternatively, University X can see if any of their students who have withdrawn enrollment ended up enrolling at another institution.

Imagine the same approach applied to online surveys. Every research company that uses panels to conduct surveys could submit a record of survey attempt/completion to the clearinghouse at the close of a project, and this record could include flags for industry-defined “bad acts.” At the onset of any panel survey, a research company could exclude bad actors from their survey knowing that they are getting complete information, spanning the whole industry (or, let’s say, ninety percent). This would be a major step forward in the reduction of survey fraud.

To raise the stakes, let’s say also that this clearinghouse tracks which panel was used in each survey instance and maintains two key metrics on all panels, visible to all research organizations. One would be the percentage of known panel members that have been flagged for fraudulent activity. The other would be the percentage of known panel members that attempt more than, say, five surveys a day.

Some might prefer such a clearinghouse to be a for-profit entity. I don’t think Rep Data, which received $6.25M in Series A funding in 2022, has the resources to purchase Imperium, nor do I think they would sell their recently-acquired Research Defender business to Dynata. The situation could always change, but that’s the way I see it now. Perhaps a coalition of survey research companies and industry associations should lead the charge.

As for the necessity of blockchain technology in this case, you would have multiple actors constantly updating a ledger of survey-taking activities, and you would need to know that the records are immutable. The panel companies being evaluated would certainly require this. Anonymity of those tagged in the system would not be a problem since you have a digital fingerprint that we will assume works adequately to identify unique survey participants. Individual panels would know the identity of their own members, but survey respondents’ actual identities could not be deduced from the data on the blockchain.

The primary barrier would be the cost and effort involved in building the blockchain platform. Considering the large amount of money wasted each year on bad survey data, I would think there is sufficient motivation to take this major step in increasing transparency and reducing survey fraud across the board.

While it may seem like it takes effort for research organizations to report their survey activity data to the clearinghouse, it only takes a couple of days, according to Research Defender’s website, to set up an initial integration between an org’s survey admin platform of choice and the fraud detection service. Once the pipeline is set up, the data will flow. Certainly, the burden is no greater than it is now.

Perhaps I’ll revisit this idea in another post later in the year. For now, if you think the clearinghouse idea has legs, be sure to comment and let other readers know and maybe send a note to your contacts at organizations like the Insights Association and ESOMAR.

As always, thanks for reading. Have a wonderful new year!

-CW