Our new GreenBook Directory site is live!
Your #1 strategic guide to consumer insights.
Qualtrics X4 The Experience Mgmt Summit
Brand & Retailer tickets for all IIeX events now start at just $99! Get or give one today!

“Push” and “Pull” Market Research

Successful market research should adopt the best of both "push marketing" and "pull marketing."



By Adriana Rocha

The business terms push and pull originated in logistics and supply chain management, but are also widely used in marketing:

Push Marketing pushes content to the consumers. Also known as “traditional marketing,” or “outbound marketing”, push is the “grandmother” of modern marketing. Direct mail marketing, such as catalogs and brochures, as well as Radio and TV Ads are prime examples of push marketing. The marketer is in control of what the message is, how it is seen, when and where.

Pull Marketing is the opposite of push marketing. Also known as “inbound marketing”, this type of marketing “pulls” a consumer into the business, meaning: the customer seeks out your company. Today’s consumer is an avid researcher. He or she reads reviews, conducts keyword searches and asks Facebook friends for suggestions. Pull marketing creates an opportunity to attract the customers who want answers you already provide. When you see a social media offer for a product you love, this is pull marketing at work. Blog posts, eBooks and other online-content machines are also forms of pull marketing that live on the web.

Let’s face it – traditional push marketing tactics are pricey and steadily becoming less effective. With an ever increasing number of ways for consumers to easily ignore advertising and find the information they want quickly online, it’s critical to understand the paradigm shift in consumer behavior that continues to rapidly proliferate: people are increasingly ignoring push marketing, and embracing inbound, or pull marketing.

What Caused the Shift from Push to Pull

More than anything, the Internet  is what has had the greatest effect. Through search engines and social networks, consumers have all of the information they need right at their fingertips, and no longer want or need unwanted ads to tell them about products or services – they find out about them on their own terms.

Until the early 2000’s when the Internet exploded into mass popularization, the main methods for businesses to market to consumers was through traditional advertising and PR. And for a long time, this worked just fine. Consumers received ads without a big fuss, and they were fairly effective at generating sales, although nearly impossible to accurately track.

But as advertising messages became more and more prevalent in virtually every aspect of daily life we as consumers have become immune to them, and started to subconsciously filter and tune out anything that smelled of advertising or sales. Then DVRs, satellite radio, email, and countless other filtering mechanisms empowered us to ignore advertising even more easily.

“Push Market Research” is in decline. 

For many years, marketing has also relied on “push market research” methodologies, in order to understand consumers’ needs, habits and behaviors. “Push market research” could be defined as a research methodology in which a marketer or researcher attempts to get their questions in front of their potential respondents, with or without them having a desire or interest to respond. Example: door-to-door, street intercept, telephone surveys, online surveys. Unless someone is passionate about answering surveys, then they probably don’t find those surveys to be entertaining or prompted at the right moment.

I’m not saying that “push market research” should be immediately considered as negative, since it can be very efficient if executed properly. However, with the quick growth of online research methodologies and proliferation of online panel companies, survey routers, etc. consumers have been bombarded with invitations to participate in surveys. People have created similar immunization to ignore survey invitations, especially because of the poor user experienced offered in most of those surveys. Additionally, consumers nowadays have many ways to contact directly the brands and organizations they want to communicate with, especially through social media. They can also express their opinions, sentiments and share experiences with other consumers, easily at any time using their mobile devices, so they don’t need to answer surveys in order to get their opinions out and to be heard.

“Pull Market Research” is on rise. 

Brands have already learned the importance of “pull market research”. Instead of just using push methods, Brands have used social media listening and advanced text analytics tools in order to understand what consumers spontaneously share in social media and public websites. However, analyzing public social media data for understanding consumers’ needs, habits and behavior has many limitations. Additionally to the lack of profiling data and superficial information available – that don’t allow researchers dig into the whys –  accessing public social media data has become increasingly challenging with social media networks such as Facebook and Twitter limiting the access to their data through public API’s.

Smart Brands have then created their own private spaces for consumers to dialogue with them directly. “Pull market research” creates an opportunity to attract the consumers who you want to talk with, empowering the users with tools to let them spontaneously express themselves, and then you can “push” questions just when needed.

I truly believe that successful market research should adopt the best of both worlds, push and pull methods, such as   insights communities – one of the fastest adopted methods by the market research industry worldwide. Companies have built research communities for many years now, but more than ever we see how important those communities have become as a source of innovation and inspiration for marketing, as well as a key part of the standard consumer insights tool-kit.

Have you combined “Push and Pull” Market Research techniques? I would love to know your thoughts and experience.


5 Differences between Push Marketing and Pull Marketing (

Push Marketing (

The Rise of Pull Marketing & Why it’s Crucial for Franchisees (

How To Balance Push And Pull Marketing(  

Please share...

22 responses to ““Push” and “Pull” Market Research

  1. Very good points. For Analytics and reporting, you then want to flip it towards a Push environment. Rather than looking for the information or the answers, they are provided to the stakeholders. Think 24/7 protocol apps, automatic data uploads, reporting and sharing, notifications and alerts etc.

  2. Recently learned about this software called qualaroo that makes it really easy to get quick feedback from a web visitor at a specific time. Seems more “push” than “pull”, but I think the line is starting to get blurred between the two as more and more of these technologies are on the rise. Thoughts?

  3. One has to be very careful about the supposed value of social media feedback for three reasons. One is most social media platforms are hardly giving any sort of representative coverage of the marketplace, if one takes out Facebook. The typical incidence of other social media media is around one in ten regular users in a major country like the UK. The second concern is that huge volumes of data on social media are basically paid for communications, in other words designed to ramp positive (or even negative!) messages. Third no one has yet worked out how to de-duplicate information on social media. In other words, how do we identify the individuality of communications that might be coming from the same user over multiple vehicles. I would add even a fourth factor, most output I have seen seems to be pretty weak data based on sentiment trends. Interest in this data starts well, but soon fades as the so called insights rarely appear and when they do raise doubts about incidence. The human overlay to make sense of that data is also huge. I doubt the efficiency of some of these so called listening tools. All good in theory, but with Push tools we at least know two things – where exactly the data comes from and its representativeness. Pull raises lots of concerns.

  4. In the context of quantitative studies, it would be quite challenging for pull MR to answer the fundamental research issues of sample representative-ness, randomization, and elimination of self-selection respondent bias. For qualitative studies, where generation of insights is the primary focus, this practice is more readily acceptable.

  5. Good article! While the concept of “pull” market research can be a little scary to market researchers as there definitely are some negatives/concerns (it doesn’t tell the complete story or the “why” behind behavior/perceptions, it’s not always representative, etc.), at the same time, it is on the rise and cannot be ignored. “Pull” market research definitely has benefits as well. For example, its speed and it can often at least uncover some hypotheses or preliminary answers to questions that can then be further explored with “push” techniques. I agree with the conclusion that market research should try to adopt the best of both worlds!

  6. great article! I believe in customers wanting to help brands with their opinion and that brands therefore should enable customers to voice their opinion and feedback when and whereever they want! So a combination of push and pull is ideal.
    Sidenote on representativeness of push: if we all agree most surveys are often long, boring and very customer unfriendly and therefore have scared away most people, why do we still think it can give us representative insights?

  7. @laura and @maaike, thanks for the feedback and comments! @maaaike that’s a great question… I believe we know that representativeness is not the ultimate goal with push methods, but even worst is continue building boring and long surveys knowing the collected data will more likely be poor and not reflect the reality or data you expect.

  8. @Chris, great points on the concerns of using social media data. Thanks for sharing the article on Marketing Week.

    @Willie, in terms of using Pull techniques on quantitative context, a great example would be to pull data from spontaneous conversations generated in a large online community (with thousand of members) and integrate with structured data collected through profiling questionnaires and surveys applied with community members. We have a great case study on this with HBO that we presented this year at ESOMAR Latin America congress. Would be happy to share more about our experience on that.

  9. Wow, there are a lot of self serving comments here about the inadequacies of push research tools. I think the time has long gone when long surveys were a technique of good market research. I question also the assumption that the survey design for these studies somehow assured poor quality data.

    These assumptions, which I think are spurious, do not provide any justification for data that lacks the basics for objective research. Essentially here I question the value of unstructured social panels collected across who knows what basis? Representation is at the heart of any scientific method.

    One may wish to claim some qualitative value from social media sample sources(I would question that based on experience!), but that may be of some use. It is intuitive in poor panel representation of anyone over 40 years and low participation rates in smartphone surveys by the same age groups. The low usage incidence of most of these media tells you something immediately! People involved in text analytic have nothing great to say about Twitter feed and most sentiment analysis is operating at about 80% accuracy – and when one looks at the data it requires a qualitative eye to make sense of it.

    Unless the target audience is some bespoke user groups, as in a community, I would have serious reservations about sampling using social media sources – and I would guess that would be the view of many good marketing directors.

  10. One other thing, there is an implicit assumption in short surveys that somehow the data can be cobbled together across multiple surveys to provide a bigger picture. This is utter nonsense as anyone knows who has been involved in trying to fuse large data sets like media data with household consumption data. Even in these data rich studies the accuracy of fused data is always a question. Anyone who claims they can link, with any objective proof. small scale surveys, is either a genius or a magician! This is akin to tales from the schoolyard. Just who do you believe?

  11. Sorry to hog the comments but I thought this was amusing and related to the discussion.

    An example of the dangers of social media listening. If even Google can get it wrong how certain are you going to be about your social media “pull” strategy? Described as an “epic” failure by Wired magazine.

    The issue always comes back to quality of the data. Seems there is a heavy human overlay required to make sense of social media. Maybe it would be quicker to just run a push survey?

  12. Interesting discussion! Overall, I don’t really think either push or pull is perfect. There really are some bad surveys out there and valid concerns about online panels (and consumers are over-surveyed). But call me old-fashioned, I still tend to lean towards Chris’ perspective that I’d trust information from a well-designed “push” survey over social media. When using “pull” techniques, I like the idea of incorporating communities, as you have a little more control over who is participating in the discussions/research. (But still, one shouldn’t throw out social media insights entirely! They’re still worth looking at…I just would be hesitant to make an important business decision based only on social media comments.)
    It’s going to be interesting to see how market research evolves, as we work to figure out how to incorporate “pull” methodologies more and develop better tools/techniques for analyzing social media and online comments.

  13. Some food for thought, Chris Robinson. Social data may not be representative, but what about the impact of its message? Isn’t there great value in knowing what others are saying because of the numbers that are listening? Example: A product on Amazon might have only 20 reviews, but how many people rely on those 20 reviews for their purchase decision? How many times have you personally relied on just one review? While social data obviously shouldn’t be used for all research objectives, I think it has great value for certain purposes. In particular, using it to understand what key influencers are saying can be a very important part of measuring perceptions and purchase drivers.

  14. Gary a good example, but note the feedback is clearly based on a real and recordable decision based on patterns of book purchasing. This is a long way away from analysis of disparate opinions which often lack any content. We all live in hope of social media being able to deliver solid content but so far the text analysis guys seem to be disappointed (without going into questions of representativeness).

    The immediacy of this information and its sheer volume suggests great potential, but it is industry dinosaurs like Laura (above) and I who ask the right question – just what does it tell us and can we trust the information? There are too many “innovative” researchers out there who need that objective viewpoint!

    Apologies Laura!!

  15. Great article Adriana!

    A lot of good things already have been said. To give my take about ”push marketresearch” since I have been in online panels for a decade, with re to the poor user experience in most online surveys: I believe the most annoying moment for a panellist is when they have to fill out a question that they already have answered when becoming a member of the panel, bad translated surveys, panellist who are being technically screened out based on their participation via mobile (phone or tablet) and so on….

    To start with fixing the online panels, I believe that online panels should work on having acurate online purchase data (some kind of a pull marketresearch) from panellist in order to validate ownership and to avoid asking the same kind of question all over again. Second, perhaps not for every survey, but we should make more efforts to give panellist the opportunity to fill out the survey via mobile (phone or tablet).

    Bottomline, we should make our surveys more fun and engaging for everyone whether pull or push!

Join the conversation