PureSpectrum - Schedule A Demo
Our new GreenBook Directory site is live!
COVID-19 guidance, tips, analysis - access full coverage here

Google Makes Their Market Research Play: Now What?

The announcement that Google had launched a formal market research offering sent shock waves through the industry today. What was surprising wasn't that they did it, but how well thought out the offering was. Is there a silver lining in this apparently oh so dark cloud?


The news shouldn’t have come to a shock to anyone;  I and others had been predicting it for years and Google had announced the Beta months ago. Nonetheless the announcement that Google had launched a formal market research offering sent shock waves through the industry today. What was surprising wasn’t that they did it but how well thought out the offering was. This isn’t half-baked; they seem in it to win it and where things go from here are interesting indeed.

This is the very definition of game changing folks. I and others have stated for years now that data  collection has become a commodity and this is the ultimate example of that; market forces dictated that this happen one way or the other.

Let’s dive into what I think this all means for our industry. First, here are the details: VentureBeat has the best coverage so far. Here  few excerpts from their article.

The quid-pro-quo system works as follows: Brands, small businesses and marketers use Google consumer surveys to conduct online market research. Google then places the questions in front of targeted web denizens when they land on a premium content page in Google’s publisher network. The brand pays per response ($0.10 to $0.50 per response depending on targeting preferences), and Google and participating publishers pocket the money made from each response. Even the would-be content viewer benefits, as she doesn’t have to pay or sign in to access news articles or videos that might otherwise be locked behind a paywall.

Here’s why this strikes us as a crafty idea: Google isn’t selling a full-service market research service, but instead a hybrid of the cheap, do-it-yourself online survey products and the expensive, hands-on approach offered by a traditional firm.

On the do-it-yourself side of things, the brand or marketer is responsible for creating the microsurvey questionnaire, determining the type of audience (U.S. only) it wants to target, and identifying the total number of responses it’s seeking. Then, the customer takes a backseat as Google drives the sampling, weighting, and analysis pieces of the process.

In so doing, Google is offering the best of both worlds — convenience, affordability, sample balancing, and analysis — and monetizing publisher content in a way that doesn’t discourage readership and traffic.

Of course, one has to wonder if Google’s controversial new privacy policies are allowing the company to pull from cross-product data to better identify folks to complete microsurveys. Google said this is not the case. The product allows for specific targeting by gender, age group, and geographic regions based on inferred demographics, and custom audience targeting through screening questions. All responses are anonymous, Google said, so they’re not tied to a user’s identity or used for future ad targeting.

It’s that last part that is most interesting, for a few reasons. Google explains their sampling model like this:

Google automatically aggregates and analyzes responses, providing the data back to you through a simple online interface. Results appear as they come in, not days or weeks later.

Results & insights

In addition to raw data, charts summarize responses and insights highlight interesting differences. Using the DoubleClick cookie and the respondent’s IP address, Google Consumer Surveys infers demographic and geographic information for each response so you can easily segment by age, gender, location and more. See which results are statistically significant or order additional responses if your initial sample wasn’t sufficient.

Methodology & accuracy

Unlike other online survey platforms which send questionnaires to predetermined “panels,” Google Consumer Surveys takes a new approach to survey sampling, data collection and post-stratification weighting. This produces a close approximation to a random sample of the US Internet population and results that are as accurate as probability based panels.

The cookie-based tracking system that underlies their sampling model is not going to fly in Europe as is, so for now the offering is U.S. centric. At first glance that may seem like a reprieve to many firms, but I think it will be only temporary, and this concern is a non-issue in developing countries so look for rapid expansion globally after the U.S. live beta.

Here is the next thing to think about: Google also owns Android and Chrome. How long do you think it will be before before this platform is deployed globally via those ecosystems? The implications for mobile sampling are particularly stunning, especially as they flood the markets in BRIC countries with low cost handsets running their platform. The incentive system can be applied to apps within the ecosystem as easily as it can be to content providers, maybe even more so. And the survey platform itself, especially the one question at a time model, is almost ideal for the mobile paradigm.

To top it off, there is the “Big Data” aspect here. Arguably, Google collects more data on consumers than any other company on the planet. The amount from search is staggering alone, but when you add in Google+, Gmail, Google Docs, Google Wallet (that one especially!) and every other offering the potential to deploy a big data driven dynamic survey system is truly within reach. Only being able to ask one question at a time may seem like a limit, but it’s not when you already have much of the data you would normally have to ask for in the traditional research model. The analytical power to deploy that may be in question today, but many companies are tackling that issue and it is likely that the combination of processing power and analytical systems needed to make this happen will be brought to market in the very short term future.

A few folks on Twitter compared this to the short lived LinkedIn research offering, but I think that it is nothing like that. LinkedIn was simply trying to sell sample with a very defined use case system in place that market researchers would not conform to. They threw their hands up in disgust and walked away because we failed to change our ways and since they didn’t control the whole value chain they couldn’t do much to influence us. Google owns the whole chain: data collection, sample, and analytics. They can and have defined the model from soup to nuts, the only question now is will there be uptake. I would argue yes, and all we need to do is look at Survey Monkey and every other self-serve player in the market. Make no mistake: clients will be flocking to this in droves because it is cheaper, faster, and easier than anything else available. To add to the value proposition, think about the almost certain integration with other data sources and Google platforms. The opportunity is simply too powerful to ignore. As Jason Anderson wrote in his prophetic post last year “we have been disintermediated”.

OK, so now what? Is there a silver lining in this apparently oh so dark cloud? Yes, I think there is and here is my take by industry segment:

  • Sample Providers: For those with a global reach, for now you have some time to drive revenue and adapt new strategies. Niche targeting, B2B, API integration with internal client systems, “Big Data” initiatives, trackers and syndicated products are all probably safe for a bit longer and smart CEOs will begin building new business models around those opportunities quickly.
  • Full Service: This wont replace trackers or more complex surveys anytime soon, nor will it replace many existing traditional projects. Much of the industry is still driven by the design & analysis anchor points of the research process and this won’t impact those offerings. Many firms are already transitioning to “strategic insight consultancies” and this should just add fuel to the fire. Ultimately it will force researchers to focus on adding real value  with our brains, not our methods.
  • Qual: This will be a boon to qualitative researchers who like a bit of quant to spice things up. I see only upside to this for now, although don’t get too comfortable since Facebook is ideally positioned to do something similar with a variety of qualitative methods.
  • Social Media and Text Analytics:  Again, only upside. For now at least Google Surveys only deal in structured survey data, although I do expect that to change. In the meantime, market researchers need to find ways to own the integration of social media and free text analysis as part of strategic insight delivery. This was a huge topic at this week’s ARF Re:think conference and the CEOs of the largest MR firms in the world all agreed with this model during the opening session with Robert Bain.
  • Specialty and niche providers. For most this will be similar to qual: just a new way to inexpensively augment their offerings.
  • DIY solutions:  It’s tougher to see the benefits to this segment. They do have entrenched users and there will be a ramping period for Google to hit their stride. Like Panel companies, this should give them time to revise their offerings and create innovative new solutions in bluer waters. What those opportunities may look like remains to be seen, but I am quite certain the entrepreneurs behind these companies will figure it out.

During the webinar on SoLoMo I moderated today with Charlie Rader of P&G, Steve Rappaport of The ARF and Andrew Jeavons of Survey Analytics we discussed how the rise of Behavioral Economics seems to have coincided with the emergence of social media/text analytics and mobile as well as the decline of traditional quant surveys. I expect to see that overarching shift continue, with more work being done to apply these models to an ever increasing tool kit of data collection techniques, including Google Surveys. Innovation and experimentation will be the new name of the game as an imperative to prosper.

Of course, I may be way off  base; only time will tell. What is unequivocal is that there is now a new major (to put it lightly) competitor within the market research space and the tension created by their entrance will create new challenges and opportunities for many. It will be interesting to watch this unfold and I for one hope that it helps to spur even more rapid innovation within our space.

Here is their marketing video: watch it and think about what I’ve said then tell me what you think.

Please share...

34 responses to “Google Makes Their Market Research Play: Now What?

  1. This is an amazing development. Imagine what better quality targets we will get from Google compared to online opt-in samples, which frankly are starting to look very weak, dubious and increasingly lacking targetability on older aged and higher-incomed segments. Hallelujah Google. Just let us know when you open this up for Asian markets. And if the industry think this will stay at a “one question only” structure you are dreaming.

  2. Feels like a game changer. There’ll be bugs I guess, but the big ‘G’ has clearly set the ball rolling for serious industry change. Fascinating blog.

  3. Great piece of work, Lenny. Fully agree with your views. I would like to add “one more thing” – market researchers can learn (from the video) how to sell themselves or to make presentations in a clear, concise and too the point manner. Thanks again for this.

  4. Apologies in advance for ‘spamming’. Please read “concise and to the point”, in lieu of too (courtesy of T9). I couldn’t resist thanking the global research community for commoditizing research. This is Google’s response to our deeds (misdeeds?) – believe me, commoditization of research knows no bounds and boundaries- kudos to globalization, in general, and multinationals, in particular!

  5. I bet the Net Promoter Score people/advocates are doing the ‘happy dance’ right now – after all, they only need one question, right?

  6. What a fantastic opportunity for Research – hitting cost, quality and speed drivers. Reminds me of poll-style singe questions that often pop-up in online Newspaper content at least in Europe. No idea how big this will be, but what an enabler! And yes, sounds like a real Game Changer to me. I love living and working in Europe but…..;)

  7. Great analysis Leonard, thank you! Fully agree with everything you say.

    Ultimately, similar to what happened with microcomputer vs IT departments and their internal clients in the late 70s/early 80s, this will allowing MR practioners and providers to focus on the more interesting and strategic and process driven aspects of MR, including how to integrate continual generation of insight data from multiple sources into the day to day decision making of product development and marketing managers.

  8. Lenny, after reading the White Paper ( ) , and reflecting a little bit more about what Google is offering with their Consumer Surveys, I start thinking their primary goal is to beat Facebook’s Questions than to compete directly with any MR player. In the past year, Facebook has promoted their “Questions” as an accurate tool for predicting public opinion. During an electoral year in the US, what best response could Google provide than launching their own questions tool, proving the results are more accuracy than any other existing probability and non-probability based Internet Surveys ?

    Well, we already knew that Google and Facebook would potentially be the next big players in the Consumer Insights arena, and here they are proving their power and capacity to innovate through state-of-the-art technologies.

    I n my opinion, Google Consumer Surveys is fascinating and well designed. However, for enterprises, it is just one more data source and tool for gathering consumer insights. It is cheap, easy to use, and seems to very accurate, but still needs to be integrated with more data, information and knowledge in order to effectively help brands and organizations take better decision making. And this is where the big opportunity for Market Researchers resides!

    It will be interesting to follow how it will evolve…

  9. What’s to make us believe the ultimate sample frame will be representative of anything-except maybe folks who are so info-starved they are willing to take surveys to be so fed? Large samples, quick data and low prices are definitely no assurance the data is reliable or valid, The sample frame must be deemed unbiased and truly representative, The “intelligence” built into the “drag and drop” platform makes me think cookie-cutter, not custom, For school groups, home owners associations and the like (think Survey Monkey), this may (?) be terrific, For decisions with significant financial consequences, my initial sense is caveat emptor, Quick n dirty and cheap isn’t where the true value in the marketing research field comes from.

    1. Great points everyone; I for one think this is a positive for the industry overall and will help many firms refocus on providing true value to clients. It won’t be great news for many others who play in the “commodity markets” of MR, but that is the nature of disruptive innovation.

      @Patrick, I don’t disagree with you, but I think it’s a mistake to think of Google as a search engine. Google has always been a data company and this is simply part of that model. They are ingenious at developing various products and services that reinforce their data assets that can be monetized as distinct revenue channels, but underlying it all is the desire to be the largest aggregator and analyzer of data in existence. The true threat to MR comes from that fact, not so much via this product alone.

  10. I don’t think it is such a bad thing personally, but even if I did, it wouldn’t matter anyway. Google are that big and dominent, they can do what they choose, when they choose.

  11. These are good points, but as a mainstream quantitative researcher I would not feel that worried. Our efficient panel providers are actually cost-competitive with Google at $0.25 per response (remember, that’s response to a question, not a whole questionnaire). For 1,000 sample that’s still $250 per question, not exactly free. And most of our surveys need to have all the questions asked of the *same* respondent so we can do relevant analysis on the data set. I’m only interested in absolute % giving a particular answer less than half the time. And what about the other limitations: can we show pictures, video, play music, track eyeballs, trade-off concept designs …

    Currently, what Google is offering seesm to have little or no relevance to the mainstream quantitative researcher. What am I missing?

    1. Good points Graham, but I think you are underestimating the power of their reach, brand, and overall value proposition. They are trying to do nothing else than disrupt the traditional research paradigm starting with marketing organizations, especially small to mid-size companies. Eventually research firms might or not embrace this model (Civic Science has been working the same idea of non-linear surveys for years, and Wal-Mart has had great success with a variation at their self-serve kiosks. Their is sufficient validation of this model for it to fly. Also, just because this is the initial offering does not mean it will always be the offering; I have every confidence that they will evolve to meet market demand, and most importantly to continue to be positioned as the central aggregation point of data; we are fooling ourselves if we many any other assumptions.

      None of this is bad news for MR; it’s just time to start thinking about the future of our industry and embracing the opportunities these new developments create for us all.

  12. @Leonard: I couldn’t agree more and I am sorry if I wasn’t clear.

    Spending the last few days at O’Reillys Where conference has been enlightening, fun and social – as most O’Reilly conferences are. Listing to the content and talking to attendees confirms one thing – There is a whole “separate” universe, trying to answer the same questions as MR but going about it a different way.

  13. A very interesting debate. However at the end of the day, the model depends on two things. 1. A person’s ability to ask the RIGHT question and 2. To APPLY the research results in existing or future marketing strategy.

    I like the idea of quick, easy, relatively cheap (and what does the Google pitch say….oh yes – “data which shows significant differences”), however as a quantitative researcher, clients (more often than not), don’t want data. They want strategy and insights, and I believe that is where MR providers need to focus and differentiate. Maybe the MR industry is a little concerned (not me though) because it means one can no longer just provide clients with a data dump.

    I strongly believe that reams of numbers in itself means relatively little, without the experience and creativity to accurately package what the numbers mean, the average Jo might not get everything they wanted from Google’s MR platform, and may wonder why their marketing strategies are not all they hoped even if they “MR’ed on Google”!

  14. I think that Google is making the right decision on developing Google as the number one company in this Industry. Keep up the good work on your Reserch on the advancement of this Company…

  15. I wonder how appropriate or valid this blog is given that you now work for the Google survey group. When was this relationship in development – that is, did you have discussions with Google at the time you wrote this article? I think it is appropriate for you to be up front about this relationship. I have tested Google’s survey capability and found it to be seriously wanting on many fronts and don’t see where the laudatory aspects based on a press release are warranted. What I’ve seen is that, in fact, this Google capability contributes to an ongoing decline in quality of internet survey research, not to an improvement. Assumed demographics? The Google whitepaper is vague at best as to how those demographics are assumed. Any mistakes in a 200 or 300 person sample could be devastating. Sounds like a return to the model of Experian’s Mosaic which assumes behavioral segments based on zip+4 and was, when I used it, quite useless for any real research purposes.

    1. Fair question John and the short answer is that I had no relationship of any kind with Google when this was written (well, other than as an Android user). It was many months later that Google approached me and asked me if I would sit on their advisory board, which is a non-compensation role and entails me (and the other members of the board) periodically giving them feedback on the development of the product and helping them explore new opportunities for application of their tools. I don’t consider this relationship to be a conflict in the slightest; the truth is that I am asked to help many firms in this way on a regular basis, some in a formal context and most in an informal way. In any event, when this was written that wasn’t even a factor; my opinion was based solely upon my observations.
      As to the product quality, I believe the overwhelming opinion by most is based on “fit for purpose”; is GCS a replacement for a conjoint study? Certainly not, nor is it designed to be. It does some things extremely well (polling, brand tracking and concept testing come to mind) and some things not so well. It’s up to the users to determine those use cases.
      Later this week I’ll be publishing a study done by Communispace testing GCS in several scenarios, as well as validating their inferred demographics. I think you’ll be impressed by what was found.
      I remain excited by the possibilities of this and many other new approaches entering the marketplace and will continue to be an advocate for experimentation and innovation across the board.

  16. Leonard – thanks for your quick reply and I’ll be giving this some thought and get back to you later on.