Return On Influence Can Return From Whence It Came

The Harvard Business Review recently published a post entitled “Return on Influence, the New ROI“. In it, the author suggested that marketers consider the use of “Return on Influence” as a metric for measuring social media activity.

What is this metric, you ask? To quote the post:

“Divide the total revenue generated via social efforts by the number of social media fans and followers, and you get a per-fan/follower value.”

There you have it – your “new” ROI – return on influence.
Really? Looks to me like that’s “Revenue Per Fan/Follower”.

Sorry, but this kind of black-hat math just doesn’t cut it. There are so many holes in the post, it’s hard to know where to begin (fortunately, Olivier Blanchard and Katie Paine did, in the comments). Still, I’ll take a stab, because I think it’s important that you, me and everyone in this space stop using BS metrics to justify social media activities and start to tie them back to business objectives.

I’ve written on this before (check out this post from two years ago) but here we go again…

1. Measuring a return requires that you compare outcomes to the input

How can you calculate a return on something without knowing what you put into it? My head hurts. This isn’t a true “return” metric; this is a poor attempt to calculate the value of a fan (without considering many of the factors in play even in that instance).

2. ROI is ROI, not Return on Imaginary Numbers

ROI has a formula. It goes like this:

(Gain from investment – cost of investment) / cost of investment

This isn’t up for negotiation. It’s a business staple. Please – if you value your job – don’t walk into a boardroom and try to sell your CFO on your fan numbers. Don’t try to sell them on retweets, or replies, or anything like that (they’re useful, but not in that context or for that audience). Show them the return that you’re able to generate for the business.

It might be hard to tie social media activities directly back to ROI, as there’s rarely a direct, solid line to be drawn (it’s extremely hard to say what, beyond the final trigger, influenced a decision to purchase, for example, but it doesn’t mean those things weren’t worthwhile). However, solid business objectives do tie back.

Which leads me to the next point…

3. Return on Influence has nothing to do with business objectives

This is something I’ve been putting a lot of thought into recently – ensuring that the social media activities we plan tie back to business objectives for our clients. Sometimes that’s sales. Sometimes that’s reduced customer churn. Sometimes it’s lowered costs.

It’s never “increasing the revenue per follower” or the “return on influence”.

 4. Measurement should be activity-specific

Imagine going pitching a metric like “return on PR”. The conversation might go something like:

You: “We calculate Return on PR by looking at the revenue generated from PR against the volume of releases we put out…”

Boss: “Get out.”

This idea is similarly ridiculous. Measure an activity, not a medium. You want to measure the ROI of a tweet? Fine. Figure out what it cost to draft/approve/publish it (time is money) and how much revenue it generated (assuming it was sales-focused). There you go – you can calculate the ROI of the tweet, and you haven’t broken a sweat yet.

Don’t try to measure the ROI of social media, or of “influence”… please.

5. Followers and fans don’t define influence

Every time someone uses reach metrics to try to define influence, a great hue and cry goes up. “It’s not reach, it’s context!” they cry. It’s true. Plus it’s a bunch of other things.

Folks like the team at Traackr have realized this, as have those at PeerIndex. Klout has cottoned-on, too, with its topic pages (although I’d still like to see them go much further down that road).

If you measure your results based on fans and followers, don’t expect senior leadership to buy into your plans for long.

6. Please – PLEASE – stop creating fake numbers

Like Ad Value Equivalency (AVE), this number tries to force a square peg through a round hole.

AVE aimed to show the value of media coverage if that same coverage had been a paid ad rather than earned media. It was bullshit, plain and simple, as it didn’t account for sentiment, credibility or any other measurement that fit around it.

There are plenty of other metrics thrown around that apply arbitrary (and, often, opaque) formulae to generate meaningless values for social media activities. I can’t stand them (plus, they violate the Barcelona Declaration of Measurement Principles, which the world’s biggest PR firms (ours included) have endorsed.

Please – let’s stop creating fake numbers and take a long, hard look instead at how we can tie our activities back to business objectives, and measure against that.


20 Responses toReturn On Influence Can Return From Whence It Came

    There goes all of my work and research into the ROIpofk.  Pronounced Roypoffck but writ as ROI for short. The Return on Interesting pictures of fluffy kittens.

    Everyone knows that each picture of a fluffy kitten generates an immediate ‘awwwwww’ of varying length and kumbayaness.  If we were to assume a base value of $0.07 per ten seconds of ‘aww’, and multiply that by the number of twitter followers, facebook fans and unique monthly visitors divided by 30, then that means this comment alone (with accompanying Ipofk) is easily worth an ROI of $1400 (where the dollar sign stands for units of awesome and not a currency).

    – – –

    Totally agree.  We have ample ways and means of measuring the outcomes of social media (or any other communications program) without resorting to a convoluted abstraction of assorted metrics.

    If you want to be taken seriously in the board room, leave existing metrics alone.  When your client is asking for an ROI, but your activities are clearly not sales driven, don’t resort to bastardizing the existing terminology by redefining it.  Point them, instead, to the proper measure.  Point them to where your activities align with business objectives and how you will demonstrate success in reaching those objectives. 

    Educate the clients – don’t bamboozle them.

    • Rob, I heart you. I also laughed out loud at your comment, leading to lots of weird looks around the office.

  • Susan Etlinger
    ago9 years

    Hear, Hear, Hear, Hear. Thank you for calling BS on this silliness.

  • Andrea Pietkiewicz
    ago9 years

    Point well made, although you just killed off the livelihoods of many “social media experts.” :p
    That said, to be able to measure the outcomes sometimes requires additional investment of budget by way of promo codes or some tactic to serve as a breadcrumb. This is something not every organization that claims it wants to measure outcomes is prepared to do. This isn’t even confined to social media.E.g. let’s say you’ve committed to sponsoring an event. How will you measure the ROI on that sponsorship? Well, you could do the lame thing and just plaster the event with logos and what-not, or you could go further and show up in person with a team of people to pitch your product, etc. Neither of these tactics results in anything that can be measured and directly attributed to any business objective.If you really want to measure the effect of the sponsorship, you’re going to have to do more: if you’re selling stuff, give a coupon with a tracking code; if you’re measuring awareness or even preference, do a short survey on-site; etc. The possibilities are endless, but they do represent additional financial investment.Those who aren’t willing to invest in tangible means of measuring outcomes might resort to these pseudo stats. To be fair, Dave, it’s often the mid-management group that gets stuck in the middle who might find these pseudo stats of interest: they can’t get the buy-in from senior management in order to get the proper budgets. It’s not an easy path they walk… #compassion

  • very nice! thanks!

  • Hi Fleet

    Great post – ROI is, as you say, in a nutshell a simple concept: what did you spend, what did you get back.
    But buried in it, are some nuances.
    For example, what counts for ‘what you get back’? One simple assumption would be to say ‘let’s look at single direct sales’ but many marketeers are more sophisticated than that, and are instead talking about CLV (customer lifetime value) a a measure of what you get back. CLV makes assumptions about how a typical customer will behave with respect to repeat purchases (assumptions= guesses, more or less informed).

    One naive assumption would be to say ‘unless we can specifically measure it, we’re not going to count it’, but that denies the impact of product contagion (and ultimately the effects of influence).
    So per several studies, we know that if 2 or 3 of my friends have bought a new product I am much more likely to convert to it (as are all the people who have 2 or 3 friends with direct experience). So what do you count in an influence based campaign? What portion of the ‘contagion / word-of-mouth’ purchases should be assigned to the influence campaign? None, all or some? And how do you measure what is a w-o-m purchase rather than a purchase driven by advertising or brand awareness?

    The critical thing marketers need are ways to reasonably and credibly assign value to these second order effects – because once they have that the ROI calculation becomes much easier.


    • @azeemazhar:disqus I would recommend – if you’ve not already read it – Don Pepper and Martha Rodger’s, Return on Customer.  An excellent read on how businesses can begin to calculate and take into consideration the long term impact they’re having on their customer base.

      I would agree that it would be naive to assume that everything can be tallied and everything counted.  And I get both the departmental and agency need to be able to claim the ‘win’ for themselves in order to be able to justify their spend, but if we hope to maintain credibility, we need to be cautious to adopt multipliers in stating our impact.  The product contagion, for example, would presumably have proportionate returns across departments and agencies.  People motivated to run out and purchase a widget(tm) because I own a widget, will do so whether I got my widget because of an editorial I read that widgets are the new black, or the catchy widget poster at my subway, or a salesman recommendation that widgets are the best hypothetical examples when I dropped by the metaphorical superstore. 

      No one department or agency is going to have enough pieces of the pie to properly infer, let alone definitively state the exact impact they have had to the bottom line.  Focus, instead, on what you know.  The outcomes that you set out to achieve.  Report on those and report on them well.  It may not be a big hairy audacious number, but it will be real.  It will be true.  And it will be something that will stand firm when giving a thorough prodding in a way that many of the multipliers do not.

      (as an aside – like the work that you’re doing on peerindex.  One of the first tools of its kind to ‘get’ that influence is contextual)

  • very interesting post! thasnks for sharing!

  • I agree that return on influence number you describe is pointless, but the reality is that all RoI measurements require you to define R and I and, in that sense, they’re all ‘imaginary’.

    How you define I – in many ways the easier of the two – depends how far you want to take the activity based costing on the time spent (and, in conventional media, whether you want to allocate out some of the time and the cost of buying the time/space whatever to other returns – does tha ad build the (book) value of the brand as well as shifting units?).

    Defining the R is equally problematic – but choosing revenue is almost as bad as trying to define Influence…

    Measurements drive behaviour, so if I’m going to be measured on revenue, I’ll just tweet a discount offer and trash the margin. And, much to our collective discomfort, sales wouldn’t stop immediately if we ceased all marketing activity tomorrow. So you have to be able to isolate the value created.Someone else mentioned Customer Lifetime Value, which is part of the answer. If you’re selling razors, selling more razors is going to increase sales of blades, for example. So your return number actually needs to be something like “lifetime contribution profit” – gross profit less allocatable costs (including other promos, back end rebates etc) plus a proportion (based on models which can be spookily accurate) of the future profit stream.Which hasn’t taken into account the profits from the sales we would have made anyway – so we have to deduct run-rate CLV. And, since we don’t want to double count, we need to add our predicted future lifetime sales (of blades) to the run-rate of those blades, so we can net them off the next blades campaign.So (incremental lifetime contribution profit – allocated cost of promotion)/allocated cost of promotion is a step in the right direction.And yes, it can be done.

  • Christ on a bike .. are we STILL talking about this? Come on people, shape up or get out.

  • The purpose of a business is to make a profit. If you’re a business then ROI is damned important.

    But if you’re not a business — or if you have enough revenue from other sources — then Return on Investment may not be the best word beginning with I, and something like Influence or Insight might work better.

  • Hello,

       I’m a civil servant in the Tourism sector here in Southern Ontario.    I work for a Destination Marketing Organization (DMO) arm of a County that serves to drive economy through tourism dollars spent across all the various towns.  My job is Content Developer for Social Media networks I manage.   I’ve learned over the past 2 years that it’s very difficult in tying dollars spent at local businesses to visits in the area.   So I’m now developing monthly, and yearly reports for my Tourism Supervisor drawing analysis insights to the impact I’m having online.  This is done using google analytics and various other reporting tools.  My next 100 miles is working on Social Media conversion insights.   I’m here looking and learning, so first thanks Dave for writing this post.  It’s funny, I had a conversation with my Supervisor over the last month.  She made us do away with using the ROI term since it’s a raw term for $ outcomes in that simple math calculation.  Then I saw Dave’s ROI “Return on Influence”.  I thought, ahhh, I have something.  Then I read the article and saw the light.  It’s not about the value of each floating scale community member, follower or fan value (Michael Slaby actually told me over coffee that would be too macro a scale to measure and not worth the time investment at SMX 2011 Toronto early this year.  That’s when I pulled back on measuring at that level. ).   I like what Iearned in this post.  It’s about tying the outcomes to goals of your organization.  So, now I’m onto something, yet what does a spreadsheet look like?  Something that has tangible meaningful metrics that correlate to my organizations goals… Hmmmm, this is my next challenge.  A report that can rank possible shares of my content, downloads of my content, positive/negative/ comments that come from my content and efforts online.  Anyone have an example document I could grow from?  @GemWebb:twitter 

  • I must say great website. I have just googled it nice info out there.

  • Thanks
    for sharing such an interesting information. 

Trackbacks & Pings