Beyond vanity metrics: How to prove true social media effectiveness
Everyone talks about social ideas, very few talk about social effectiveness. This, as is often the case with the shiny versus the significant, is a mistake. Now, this isn’t a dig at social ideas. But ideas, I’m afraid, are only as good as the story you can tell about them. Not to people outside of a company. But to the people who operate inside it.
Mastering effectiveness stories is the act of being able to communicate internally not just what an insight was, or a strategy, or an idea, but the impact of those things combined.
Here’s the problem, though. You might think you have a social effectiveness story when you talk about reach, or video views, or engagements, or even clicks, but that’s only part of it.
Why your social media ROI reports only tell 33% of the story
In fact, 15 years in this game have taught me that most social reports only tell about 33% of a story. Because the above metrics cover the impact a piece of work had on channels or media results, but that’s only one part of it.
The other two parts are how it’s affected the brand, and how it’s affected the commercial results of a business. Or at least proxies for these things.
If you want a framework to capture all of it, it looks like this:
| Measurement Tier | What It Actually Proves | Metrics From This Guide |
|---|---|---|
| Level 1: Media Effects | Channel baseline & exposure | Reach, Impressions / CPM, Video views / VTR, Engagements / ER, Clicks / CTR |
| Level 2: Media + Creative | Message resonance & execution | Ad recall, Consideration lift, Message lift, Associations lift, Intent lift |
| Level 3: Brand Impact | Long-term brand health & growth | Brand awareness, Consideration, Affinity, Image, Intent |
| Level 9,000: Business Impact | Commercial return on investment | Total volume orders, Cost Per Acquisition (CPA), Penetration, Market share, Net profit |
The 3 levels of social media measurement (and the Holy Grail)
Level 1: Media effects (the baseline)
The standard stuff you and I report on. Essentially, by producing a certain amount of pieces of content, we delivered better channel performance. This could happen at an owned, paid or earned level. It’s where most people will report on stuff, but it’s often point of parity metrics that don’t mean much:
Reach & Impressions / CPM
Video views / VTR
Engagements / ER
Clicks / CTR
Level 2: Media + creative (the lift)
Then, if you’re feeling extra fancy, you may even design your measurement studies to include stuff like:
Ad recall
Consideration lift
Message lift
Associations lift
Intent lift
Which essentially tell you that a particular execution, or set of executions, have done the creative job they set out to do.
Harder to achieve, but possible if you buy brand lift studies. Or if you do control vs exposed testing, or real world testing, with platforms like ThisThat or by running targeted surveys.
The Point: You need a measure of what people took out of all that exposure and engagement. Without this, you don’t have much to compare social activity to wider brand activity, which matters for a growing number of businesses.
Level 3: Brand impact (the long game)
Now we start getting serious. We’re not just talking about impressions, or volume of engagement (which, by the way, has no correlation with brand effects, though can help with some baseline salience). We’re talking beyond the impression you left with people, not just how many impressions you got.
This is also where it gets trickier because there are so many factors that affect the health of a brand, but you want to be able to map some of your work against growth in areas like:
Brand awareness
Consideration
Affinity
Image
Intent
The crucial thing here is we’re not just looking at it on a campaign by campaign basis, but more as something you can track over time. So yeah, you’ll need to have this set up for the longer term, make social be part of that conversation somehow, stay distinctive and consistent.
(If it were easy everyone would do it right?)
Level 9,000: Business impact (the Holy Grail)
The holy grail if you can get to this as well as the previous two. Admittedly, this is the tough one to track, but there are a few possibilities to help you paint a richer picture.
The ideal route (econometrics): To invest in marketing mix modelling, but let’s be real: your social spend is likely not going to be sufficient when compared to wider marketing activities and macro factors let alone advertising, to show a significant dent. The best case I’ve seen this done with was Cadbury when I worked with them at VCCP, and it was based on five or six-digit investments in paid social, consistently, across platforms. So, top 1% stuff really. Not (yet) for most brands.
The pragmatic Route (correlation analysis): Imperfect, no doubt (correlation is not causation), but better than nothing. Here, you’re trying to determine patterns between your overall social channel performance, and any commercial results that may have occurred some time later. This is also heavily category dependent, as the purchase cycle for chocolate will be shorter than, say, a mobile phone.
Social media effectiveness examples done right
Speaking of which, a few great templates of how to think about social (and all) effectiveness include:
O2 and Samsung (UK): We were able to show that by priming people with influencer content before re-targeting them with performance ads, we were able to increase total volume orders while having a much lower cost per acquisition. This led to the campaign becoming the template for how to do future device-specific communications on social media.
Ryanair: Under Michael Corcoran, they combined channel metrics with post-purchase surveys, essentially creating a more causal link between millions of impressions on social, and people who, after a purchase, said on a survey that they were last exposed to Ryanair communications on social. Again, not perfect, but better than nothing. Something is always better than nothing, and the more ‘somethings’ you have to tell the story, the better you end up in the end.
Aldi (UK): Their IPA effectiveness paper shows an extremely strong case for their reactive and planned reactive social presence generating impressive results.
Channel Impact: A 64% increase in reach and 91% in engagement between 2020 and 2023.
Brand Impact: Out-performed buzz scores by triple percentage digits vs Asda, Sainsbury’s and Tesco and increased 19% in brand salience.
Business Impact: 51% penetration among pre-families, a 10.2% market share in summer 2023 (vs 7.1% in January 2021), and nearly £42 million in net profit over 3 years. Again, big league stuff, but this is what we should try and aim for eventually.
Approach data like a lawyer, not a scientist
What’s common with these examples is a change in mentality: you’re not making a scientific argument in a lab. You’re making a case in court. So the more evidence you can gather, and the more coherent your story can be based on that evidence, the better you will do.
For this to be effective you need by definition multiple sources. Some operating assumptions behind how those sources tell a clear enough story. But my hope is the above set of tools and techniques can help with a variety of budgets.
All of this still requires some money and time, though:
On Money: A generally wise rule of thumb is to allocate 5–15% of your production spend to measurement.
On Time: Be aware that most long-term brand metrics are best measured through a tracker, over more than just a few months; some of the most effective work ever tracks performance over a 3+ year period, or longer.
But again, something is better than nothing, and having more than channel metrics will already help set you apart from 90% of social pros, who still insist on engagement rates or sentiment scores based on 11 random comments on a defunct trend’s TikTok.
Contextualise everything
A couple final pieces of advice to make these results mean something:
Ensure you have a clear basis for comparison: Own past results, industry benchmarks, platform benchmarks. Otherwise, it’s impossible to know whether 10 million of something is good, bad or ugly. Many people miss this.
Define measures and targets before briefing the work: This will ensure everyone is aware of how success will be evaluated. We rarely do this, because we just want to see ideas, but this is the difference between having the ‘boring’ conversation now and a painful one in 6 months’ time. I’ll take boring over painful any day of the week, or quarter of the year.
Once you’ve had this conversation, the fun part can begin. But the fun part ain’t fun for long until you sort the fundamentals that will help you keep pitching for budget, and getting it because you’ve proven there’s something we gain from the work.
And that’s the difference between how a proper social pro thinks about effectiveness with a wider view, and someone who decided to ‘do some socials’ because it felt like a fun way to spend the workdays.
Want more like this? Subscribe here:
