Calculating ROI from Social Media - Problems, Pitfalls & Breaking all the things...

As an industry we're quite rightly under pressure to report and measure the impact of our activity in order to justify our existence.

How to calculate return on investment (ROI) on social media activity is something which I'm asked a lot - I guess because ROI is a metric that the C-Suite are both familiar with and very interested in.

However, ROI is a woefully poor measure of the success or failure of social media activity.


To be clear, I'm not saying ROI is a bad metric. But it's misunderstood and often misappropriated.

But I'm leaping ahead of myself. First let's explore the metric.


How to calculate Return on Investment (ROI)

If you want to calculate ROI you'll need to understand both the value of what you've invested (in money) and also the the value of what you've got back (in money).

Here's how you calculate ROI:

(gained from investment - cost of investment) / cost of investment

Let's do a quick worked example:

(gained from investment - cost of investment) / cost of investment

(£1000 - £100) / £100 = ROI of 9

NB sometimes ROI is expressed as a percentage - should you wish to express is as such simply multiply by 100 - in this instance the above example would yield a ROI of 900%.


So what are the problems with ROI as a metric?

ROI as a metric is only useful (and indeed should only be used) when looking to understand how cost-effectively various tactics have performed historically.

I strongly dislike talking in the abstract so I'm using an example:

Tactic ROI
Email 50
Social Media 2





Taking the example above you can see that email is clearly the most cost-effective tactic, generating £50 for every £1 spent.

However, it often doesn't stop there. Rather than simply looking to see which tactics are most cost-effective there's often a temptation to use ROI to determine where to invest marketing budget in the future.

That's where it can get messy.

As soon as you use ROI to determine where to invest you might be in danger of making the assumption that costs scale with success.

In almost all cases this simply isn't true.

Using the example above let's imagine that the company invested £10k in email marketing to generate £500k. Lovely.

But does it follow that if they invest £20k next year, they'll make £1m?

Actually, no - it doesn't. If you up-scale email activity without due care and attention you might find that instead of making lots more lovely money, you instead annoy your subscribers (perhaps to the extent that they simply unsubscribe). It's a delicate balancing act.

You'll find similar issues with other tactics too. Upscaling PPC budgets might not automatically scale either. If you have a very small PPC budget it's likely that you're focusing on a core set of very relevant terms, hitting the long tail hard because the clicks are cheaper etc. But that is a limited universe. At some point you'll have covered all of the cheapest CPC terms and in order to spend more budget at some point you'll need to go after the more expensive CPC terms. At that point you'll typically see ROI drop.

The truth is that the vast majority of activities simply do not scale up and retain the same ROI.


Furthermore, ROI neatly masks the bigger picture - actual gains. In the example above, email looks like an amazing profit driver. But is that actually the truth? You don't know. You can't see that from this metric. All you can see is how cost-effective each tactic is - not how much actual money you're making.

Of course businesses want to use cost effective marketing tactics, but ultimately they really need to hit their financial targets - i.e. make money.


ROI as a metric masks how much actual money has been made

As a business would you rather spend £10 and make £1k (ROI of 100), or spend £100k and make £1m (ROI of 10)?

Sure, spending just £10 and making £1k sounds really great, but if your business makes a total of £1k per year frankly it's not really a business it's a hobby - it can't even keep a roof over your head. At some point (in order to grow from a hobby to an actual business) you're going to need to accept lower ROI activity.


OK so I've explored some of the problems with ROI as a metric. Let's assume you accept the limitations of the metric but nevertheless want (or need) to calculate ROI.

You'll need to understand both the value of what you've invested (in money) and also the the value of what you've got back (in money).


Calculating this stuff is hard.

Celeste Behind the Abacus

In many directions. First off it might be difficult to determine what you've spent on social media... What has it cost you?

Lots of (wo)man hours, perhaps you're paying an agency, perhaps you're paying for social seeding - also what about your content? How much does that cost? More (wo)man hours? Another agency? Is it just for social media? Or are there other channels that benefit from the content? How do you split the cost of that content out?

But that's a cake walk compared to determining what you've gained from your investment.


Strictly speaking you should only include the monetary value of what this activity has yielded. So perhaps you'll look at conversions generated directly from social media, or perhaps you'll look at assisted conversions, or multi-touch attribution.


Trouble is, those stats don't look great

Social media isn't the best way of increasing revenue; at least not directly. If your primary objective for your social media activity is revenue driven I'd suggest that you're likely to fail. It's not that social media can't lead to revenue. It's just not the best tool to generate direct revenue.

If it's direct sales that you want, something like PPC is likely to yield a better ROI.

Why? Well (assuming you've set up your PPC campaigns properly) you're popping your sales messages right in front of consumers who are actively seeking your product or service. Those consumers are in a researching or even buying mindset.

Are those consumers on Twitter, Facebook et al also in a researching or buying mindset? Not so much. I can't remember the last time I thought to myself - I really need to renew my home insurance - better head on over to Facebook and do that.

And there's the rub. You are not comparing eggs with eggs. You are comparing consumers in vastly different stages of the purchase funnel. Social Media doesn't deliver very impressive stats if you're just looking at these revenue metrics.

We need Social Media to show a more impressive ROI, so what do we do? We start trying to put £ values on some of the other metrics which we're measuring.


And so it all goes to hell...

What is a twitter follower worth? Or a RT? Someone 'encircling' you (still sounds creepy) on G+? What about a Facebook like? How do you go about sticking a £ value on that?


Mis-calculating what a 'like' is worth

Every time some lazy journalist posts nonsense about a Facebook 'like' being worth $174.17 I die, a little.

OK, clearly I'm dramatising - but seriously?

The problem is we're conflating things here.


It doesn't surprise me that consumers who like a brand enough to 'like' a brand's Facebook page would spend more money with said brand. It's likely they liked the page because they like the brand.

No, I mean really actually like the brand. Not just Facebook 'liked' the brand. Aaaarrggghhh this is getting confusing...


Let's try with an example:

Phil really likes cables. Audio cables. Stuff like that. So much so he 'likes' the page of his favourite audio cable supplier - Ron's Really Rather Lovely Audio Cable Shack.

I on the other hand do not like cables. I have never purchased a cable. All the stuff I buy already comes with cables. Why would I want to buy cables? As such, despite seeing that Phil has liked Ron's Really Rather Lovely Audio Cable Shack I still elect not to 'like' said page.

At the end of the year Phil has spent many hundreds of dollars with Ron's Really Rather Lovely Audio Cable Shack. I have not.

However that's got nothing to do with Facebook likes - that's got to do with real world likes.


It's very difficult to place a reliable revenue value on any of these sorts of metrics. Are your Facebook fans spending more with you than non-Facebook fans just because they love you? Or is their engagement on Facebook actually driving more purchases?

Even if you manage to figure this stuff out are you double-counting this stuff? If a Facebook fan buys as a result of a Google organic click are you attributing that sale to Facebook *and* SEO?



We are in danger of breaking things


Social media is not a sales channel in the traditional sense. You can't just broadcast your messages and expect to yield sales in return. That's just not what it's good for. 

Analysing the efficacy of our activities on social media by looking at metrics like ROI might well push us towards the wrong sorts of activity - i.e. in a bid to increase ROI from the channel you may be tempted to release a relentless swathe of sales messages. But this is likely to turn your audience off and than rather than driving sales up you may find that you're driving your audience away.


So what is Social Media good for?

I think good uses of social media include:
  • Getting people talking about you (and indeed to you) and/or shifting consumer perceptions of your brand
  • Customer service / reputation management
These two very distinct uses of social media require two distinct sets of metrics to analyse success.


Getting people talking about you

If you're engaging in social media because you want to get people talking about you and/or to shift consumer opinion, your ultimate goal ought to be to grow an engaged and relevant following.

As such you should be measuring things which will help you to understand to what degree you've succeeded.

'Good' metrics:

  • Engagement / Advocacy
    • Fans and followers should be tracked over time, but this is a potentially deceptive metric. If people follow you but then never interact with you again then you've not built an engaged following.
    • Focus on interactions - retweets / @ mentions / posts / comments. See which content yields the most interactions (and the least) and do more of what works best for your audience.
  • Brand Awareness / Brand Perception
    • If you're seeking to change brand perception or increase brand awareness you're going to need to measure it frequently (e.g. once per quarter) to understand if things are moving in the right direction. I'd recommend using a market research agency to do so. However, if this is outside of your budget and you elect to conduct this research in-house then take the time to understand how to properly conduct sound brand awareness / brand perception research - i.e. don't just poll your Facebook fans to see what they think - you're in danger of getting a biased picture.
    • You could also benchmark (and then measure) branded traffic (might be tricky thanks to not provided) and direct traffic to your site. However, be careful here. Whilst your activity via social media may well have increased brand awareness and/or improved brand perception so that people are now coming direct to your site, can this solely be attributed to your activity? Or was other activity also being undertaken during this time? If you're not 100% sure that the change is down to your activity don't claim 100% of the uplift.
  • PR Value
    • Has your social media activity attracted coverage outside of the social media space? Or has your activity led to other opportunities? E.G. - have you been invited to speak at a conference, contribute to a study, be interviewed, comment on a new story etc? If so you might elect to use a 'advertising value equivalency' to try to put a value on the coverage. (NB there are pitfalls here too - but if your organisation understands this metric go ahead and speak their language). Alternatively you could simply record / monitor the coverage.
For what it's worth I would still also measure both raw visits from social media, plus conversions and assisted conversions. However they should not be the only thing you report on.


Customer Service / Reputation Management

If you're using social media as a customer service channel then I'd argue that ROI is potentially a pretty redundant metric. I think it's fair to say that most Customer Service departments don't report on the ROI of their activity.

Instead I would advocate the use of traditional customer service metrics like:

  • Number of queries dealt with
  • Number of queries dealt with without escalation
  • Time to resolution

If you also have traditional customer service departments you might elect to use the following formula (this is what the team at Xbox use):

Number of unique customers engaged with Xbox via twitter

(multiplied by)

% of people who say they would have called instead of tweeting

(multiplied by)

Average cost per call

= Money saved in call centre costs


Again here, for what it's worth I'm not sure quite how much I like this metric as I think it still has the potential to be misleading, however if the C-Suite are signed up to it, then by all means use it.


I also wanted to make mention of reputation management here. Again, like customer service I don't think you can accurately hope to measure the ROI of reputation management.

How much might averting a PR disaster be worth to your company? That is very hard to put a figure on. Nevertheless it is important. As such I would monitor and report on all reputation management activity which is undertaken via social media.


Let's round this thing up...

To conclude, when it comes to calculating ROI from Social Media I'm concerned on a couple of counts:
  • ROI as a metric is problematic - whilst is does a good job of telling you 
  • how cost effective one particular tactic is versus another:
    • It masks how much revenue has actually been made
    • There may be a tendency to assume that you can scale up activity and retain the same ROI - but this often isn't the case
  • Social media doesn't typically generate fantastic ROI versus other channels.
    • Direct and assisted conversions via social media are typically lower. This is at least in part down to consumers on social networks not necessarily being in the same stage of the purchase funnel as those who are clicking through PPC ads / clicking through organic listings.
  • In an attempt to increase perceived ROI (and justify our existence) we try to put monetary values on things like Twitter followers / Facebook likes etc
    • But our methods for doing so are often flawed and our data is not trustworthy.

If you accept that social media is about conversation not broadcast, and also accept that 'good' uses of social media look like this:

  • Getting people talking about you (and indeed to you) and/or shifting consumer perceptions of your brand
  • Customer service / reputation management
... Then surely you must also accept that ROI is a woefully poor metric to measure either.


Rather than coming up with more fancy ways to calculate the incalculable I think we need to accept that social media just doesn't deliver great ROI directly.

But that's OK.

We need to acknowledge that ROI simply isn't the right metric to use and instead measure things which accurately reflect what we're actually looking to achieve via social media.

Just because an activity doesn't have tangible ROI, doesn't mean that it's not valuable or useful.


And so dear reader, over to you - agree? Disagree? I'd love to hear your thoughts via the comments.


Image credits:

Droids, Abacus, Breaking Things


Get blog posts via email