Forrester: Facebook and Twitter Do Almost Nothing for Sales

Is that really the case?

This headline from Business Insider has to be the most misleading of the year for me. It’s certainly engaging, but what's it based upon? When Andy Lloyd Gordon shared it through Twitter yesterday I couldn’t resist clicking to see what this claim was based on.

The report by Sucharita Mulpuru, an analyst at Forrester Research showed that less than 1 percent of the online transactions amongst US customers could be traced to a social media post.



The article states:

"Social tactics are not meaningful sales drivers.

While the hype around social networks as a driver of influence in eCommerce continues to capture the attention of online executives, the truth is that social continues to struggle and registers as a barely negligible source of sales for either new or repeat buyers.

In fact, fewer than 1% of transactions for both new and repeat shoppers could be traced back to trackable social links."

Retail transactions from 2012 included single touchpoint transactions and multiple touchpoint attribution. For new customers acquisition, 67% of which involved a single touchpoint, 20% were direct visits, 16% were organic search and 11% were paid search. Social media were less than 1% for both single touchpoint and multiple touchpoint attribution analysis.


So what is this based on and who does it refer to? Well it isn’t a survey of retailer or consumer perceptions. Business Insider didn’t go into the methodology, that might get in the way of the headline, but it’s based on primary web analytics data of transactions from US customers of retail clients of GSI Ecommerce.

The clients of GSI are well-known retail brands particularly in clothing, e.g. DKNY, Levis and Speedo and toys (ToysRus), Early Learning Centre and Mattel.

So, it’s a credible source, indeed we featured data from the same survey a year ago when discussing the importance of Email marketing to retail sales - that’s the angle we took…

Business Insider noted:

Mulpuru didn’t study small businesses, which she said do disproportionately well in social commerce”.

Despite this, even for big brands we might expect to see more direct influence of social on sales since this is not a “last click” study, the research does look at multiple touchpoints although the cookie window for attribution isn’t reviewed, we can assume it’s 30 days?

So what can we learn from this given that this is a credible study? Well it’s certainly a reminder of the power of other digital channels in driving sales. Paid and Organic, email marketing (from existing customers) and to a lesser extent display advertising and comparison sites are undeniably still key sales drivers so they merit investment and work on improving their efficiency through following best practices.

Why this headline and data gives a misleading view of the value of social media marketing?

I still think that the conclusion that Facebook and Twitter Do almost nothing to drive sales is still highly misleading, since a direct response view of effectiveness. For me, investment in engagement through social media is still warranted since this study doesn’t look at what develops brand awareness and brand preference at the top of the conversion funnel.

It’s also worth remembering the difficulties of tracking social media effectiveness. Indeed many of the visits from direct site visits (the top referrer at 20%) and brand search terms may have been influenced by social media engagement. Then there's the role of social media in influencing offline sales.

There are also technical reasons why social media may be undercounted including:

  • Social media may not be tracked from mobile and desktop apps and would show as “direct visits”
  • Users of social media will not necessarily click on the links within social media, but will search or to direct to the sites when interested in purchase
  • For a large brand a significant proportion of search traffic will be brand terms showing there is already existing awareness and preference for a brand. This may have been generated through social media and can also increase conversion rates.
  • Returning visitors with multiple visits, some of which may be referred by social media may not be tracked since cookies are deleted, the cookie tracking window is limited or they have visited on multiple devices.

What do you think? Do you think this report suggests we may be paying too much attention and focusing too much resource on social media or do you think the report is misleading?

Share your thoughts

  • Should this really surprise us? Isn’t Facebook and all social not really about the sales but being social with our already established customers and making them advocates for selling to others and keeping our brands near their hearts rather than instantly opening up their wallets. Money is very important – but not the point of social.

    The headline is like one saying no business comes from networking events or out of drinking in bars!

    Most transactions aren’t instant. So “business” might not happen right there and then (or if it does it is very rare to and different depending on b2b and b2c) but it happens later on and with social disproportionately well for SME’s 😉

    Great end point. It might be the way it’s tracked after all of this…..

    • Hello Dan, you’re right! This is why I disliked the over-simplistic implication of the BI piece that “Social media are pointless”.
      They should have commented in a more balanced article, as you have:

      “Most transactions aren’t instant. So “business” might not happen right there and then (or if it does it is very rare to and different depending on b2b and b2c) but it happens later on and with social disproportionately well for SME’s ;)”.
      You need to develop brand awareness and preference to help convert online visitors and this benefit of social media marketing isn’t being measured here.

  • adair2 commented on November 23, 2012

    So we could say, paraphrasing a famous “old media” quip, I know that half of my social media marketing is working, but for technical reasons, I don’t know which half.

    • nice – wish i had thought of that 🙂

    • Nice!

      But do we even know that half is working? You’re right to highlight the technical limitations – I fear we will never know which half… You can’t really do controlled testing of social media like paid media unfortunately.

  • Maria Morais commented on November 22, 2012

    Hi Dave,

    Congratulations on your ‘maybes’ and hypotheses, this is how a social scientist creates assumptions (starting point for any research). Social media is not an exact science so it can’t be treated like one. This morning I read an amazing post from Patricio Robles at Econsultancy about Amazon going social. Would be interesting to hear your thoughts on this.

    • Hi Maria

      Surely you’re having your cake and eating it?

      Of course, ‘maybes’ and hypotheses are the starting point for any research (including social science).

      However, if we’re going to label our opinions as hypotheses without testing them, we’re simply wearing the cloak of analytic respectability and business responsibility without doing the work (or the math).

      For a hypothesis to be of any use, it needs to be either proved or disproved.

      And, unless you’re a physicist or other ‘proper scientist’ (who may dismiss the social sciences as unscientific), social science can go on to test its hypotheses. Visit any university sociology or psychology department and you’ll see what I mean.

      My logic is (and correct me if I’m wrong please), if social media is a human activity, and human activity can be observed, measured and researched by the social sciences (in a scientific manner) then it too can be treated scientifically.

      And if social media can be described as a digital marketing activity, then it is even more open to research, tracking and analysis.

      Do you agree?

      • Maria Morais commented on November 22, 2012

        Hi Andrew,

        I couldn’t agree more with you. Of course that social sciences can be treated scientifically and they are, even for ‘proper scientists’ as you call them. This discussion about what is science was clearly not my point, maybe I was not clear in my comment and I am sorry for that.

        My logic is, social media is about people and behaviours, so sales shouldn’t be the major topic when evaluating Facebook and Twitter. People behave beyond economic factors.

        Bottom line, with all the respect I have for Forrester Research, the focus on sales isn’t the right angle to evaluate social media, but this is only my opinion, no science and no cakes behind it 🙂

        • Thanks for your comments and clarification Maria, this is why I was criticising the simplistic write up on BI.
          Engagement through social media can help generate awareness and preference for brands which ultimately do impact sales and this significant benefit of social media was not covered in the Business Insider posts – it’s more their write up than the Forrester research I was calling out.

        • Hi Maria

          Thanks for clarifying.

          My reference to social science was that, if we’re going to describe statements as ‘hypotheses’ and use the labels of science, then we should be willing to test them.

          If we don’t, then they’re not really hypotheses. They’re simply guesses or hunches (as you’re graciously pointing out).

          For example, if you work with hard pressed, small organisations who have limited marketing budgets (as I often do) then why shouldn’t sales not be a legitimate metric for their social media activity?

          It may not be the only goal for them but for these organisations (and many others) sales attributed to social media is a major topic!

          To put it another way, if sales isn’t ‘the right angle’ to evaluate social media (at least for commercial organisations, promoting and selling goods and services) then what is?

          • Maria Morais commented on November 22, 2012

            It’s a fair question but I guess there’s not a lot I can add now. My point is that consumers are using social media for lifestyle and entertainment reasons before they think about buying products and there are countless studies that indicate that. But yes I understand that some organisations don’t see any advantage in establish a conversation with followers when by the end of the day they will be measured by results.
            I am going to read the source when published by GSI. Thanks for this conversation and insights.

          • Hi Maria

            Using Social Media for lifestyle and conversational reasons is perfectly valid. I do so myself all of the time.

            Establishing a conversation with followers is valid for every organisation.

            But if you’ve got a limited budget and finite resources, you can’t – and shouldn’t – do everything. You have to prioritise and know what works in some sort of priority order.

            Which is why, if the GSI research indicates social plays little part in direct sales – although it probably plays some – is the reason we all need to know.

            I’ll be interested in your thoughts when you’ve read the source material.

            And thanks to you too for your time taking part in this conversation 🙂

  • Hi Dave

    Thanks for referencing me in the article. Although I note you don’t mention others who tweeted it too. It’s resonated with lots of people.

    Seeing as you have linked me with something you find controversial and misleading (I didn’t write the report you know!) then I thought I’d make a contribution.

    Firstly, regarding the headline on the Business Insider (BI) piece. I understand why you think it might be misleading. But I’d suggest that it’s a reference to the research they’re reporting on i.e. ‘social (in Forrester’s survey) didn’t drive sales’.

    A topical, controversial ‘make ’em wanna’ read’ headline maybe. But misleading? I don’t know. Have you pointed out your concerns to BI on their website?

    Another question – have you read the original research (I sent you a link in a tweet to the Forrester page)? Or is your article simply commenting on the Business Insider piece i.e. have you read the source material?

    I haven’t. And, as is the case on Twitter, I was simply sharing an article I thought interesting.

    Thus, unless we’ve read the source material we – but especially me – need to be careful about making assumptions on data collection, methodology and so on. Indeed, the source material may or may not reveal these in full detail. If it doesn’t, then we’d have to go back to the report’s author for clarification.

    With these caveats in mind, I must admit to being surprised that social media does not, apparently, drive sales. In this study at least. Especially when the focus of the report is on big brands that have, usually, big social media budgets, large social media teams and footprints.

    Maybe, after all, social media doesn’t generate sales? Shock! Horror!

    However, we should be delighted there appears to be a ‘proper’ piece of social media research done by a ‘proper’ research organisation. For too long, social media’s contribution to the bottom line has been (still is often) based on an intuitive, anecdotal, almost gut-instinct basis.

    Indeed, your blog post is also full of ‘maybe’s’. Such as in ‘social media may not be tracked from mobile’ or ‘returning visits referred by social media may not be tracked’ etc. Maybe you’re right. Maybe you’re wrong.

    But with so much time, energy and money invested in social media marketing, we can’t go on with ‘maybe’s’ anymore. We need evidence. One way or another.

    If this report counters our intuitive views, then great! That’s how any proper, mature industry moves forward. We might suffer from what psychologists call ‘cognitive dissonance’ (when we want to dismiss or ignore what’s being presented to us). But as professionals we should be happy that the field has moved forward a step.

    Whilst our fluffy marketing activities can rarely be viewed as ‘scientific’, we should try and learn from the scientific method.

    If a piece of research surprises us then that’s good. If a piece of research runs counter to our world-view then good. If we question a piece of research, it’s methods, analysis and conclusions then, good.

    We can, and should, question the findings of any research (as you’ve done).

    But if we want to show that the research is incorrect, or indeed, misleading, then we should point to similar research that counters it. And, if we really want to emulate science, we should also run our own piece of research that demonstrates a different view. Because, until that time, we’re simply operating on ‘maybe’s’.

    Now…back to my Twitter, LinkedIn pages, Facebook work, Instagram profile, YouTube channel etc etc 😉

    • Thanks for your extended response – this is a BIG issue for marketers which is why I wanted to call out the research, or rather the way it was reported.

      Yes, it was good to surprise and I’d hope people do question, but many don’t have time – images or infographics can often mislead since they don’t give the full picture and hardly ever mention methodology, sample size, frame, yet still they influence. I always try to report on these on this site.

      I have used “maybes” because I’m a scientist by training and I dislike definitive statements as in the BI post unless they can be backed up.

      We see a lot of these sensationalist headlines in social media since they engage and they’re shared – I’m guilty of this myself. Lazy headlines do influence people, like the “Pinterest is more popular than XYZ combined” used to refer to one site.

      Have a good day!

      • Hi Dave

        Thanks for taking the time to respond.

        I’m a little confused on whether you’ve read the original research or not? And, therefore, whether your thoughts (as in your blog post) are based on your analysis of the research itself?

        Or, whether your piece is a response just to the reporting of the research (on the BI site)?

        As I say, I don’t feel I can comment properly on the findings of the Forrester stuff because I’ve not read the original research (as in the second link I tweeted you).

        I recognise your substantial scientific training.

        However, as I was pointing out, in healthy scientific discourse, shouldn’t we (ideally) go back to the source material?

        Only then can we truly question the research methodology and, if we need to, point out its flaws and weaknesses. We can then, perhaps, point to other – soundly conducted – research that supports our alternative assertions/hypotheses. And so the cycle continues 🙂



        • Thanks for the probing Andrew!

          Of course, you’re correct about consulting the source, but my point is that most readers won’t, so as a minimum we should question the methodology and state it which I have.
          I read the equivalent full report in 2011 published by GSI and written up by Forrester, but have not read the 2012 report since it’s $500 – I’m waiting for it to be released via GSI – I checked and they haven’t released it yet. From knowing the previous research and reading the Forrester blog, I felt I could give more info on methodology than the Business Insider post did and question their definitive statements.
          I’m not “bigging up” my “scientific training” in fact it’s seriously dated since it was in the 1980s, just reinforcing what Maria was suggesting that we should try to evaluate reports more rigorously. This is difficult if the methodology isn’t reported.
          As I said in my post, what’s most interesting to me, is the implications of the research and how it’s presented on marketing in the real-world – i.e. the resource/focus that marketers put into social media against other digital media channels. This report should certainly make us question that…


          • Social media is good for probing Dave!

            The point of my responses are, if we are going to criticise methodologies and research (rather than simply report it as I did in my tweet) then I feel we can only do so properly when we’ve read and checked the source material.

            Otherwise, what we’re doing (and are doing right now) is simply criticising the reporting of the research. A debate that may, or may not have any validity or usefulness!

            By the way, you should ‘big up’ your scientific training. It’s better than mine 🙂



          • Let’s leave it there then Andrew – that was one of my aims – to criticise the reporting of the research while also prompt questioning of how marketers value the contribution from social media and the amount of time they put into it.
            Thanks for the conversation!

Get FREE marketing planning templates

Start your Digital Marketing Plan today with our free Basic membership.

  • FREE fast start guides to review your approach
  • FREE digital marketing plan templates
  • FREE alerts on the latest developments

Need help with your Social media marketing?

Get more from your digital marketing with in-company or remote training and consultancy from Smart Insights

Get FREE marketing planning templates

Start your Digital Marketing Plan today with our free Basic membership.

  • FREE fast start guides to review your approach
  • FREE digital marketing plan templates
  • FREE alerts on the latest developments