Discovering the Unexpected with Paid Social — and why Facebook’s current best-practice guidelines don’t make sense to me.
For an advertiser, advertising via Paid Social is manna from heaven.
As an advertiser, I can draw upon a plethora of different formats, from ones that allow me to showcase my wares, update my pricing or actively drive leads there and then.
I can quickly and easily reach an audience of prospective customers by targeting them based on what the social media platform knows about their demography, interests and behaviours.
If I want to target millennials, I go get them. Middle aged women? They’re out there. People in market for a car? Or accounting services for their small business? No problem.
However, as a Paid Media consultant, I am frequently brought into projects with a view to auditing their performance so as to then make recommendations on where to go next. One thing crops up fairly consistently…
Advertisers tend not to split their campaigns down far enough.
And there’s gold in them there hills!
Let me explain: when thinking of whom to target with a Paid Social campaign, there’s a variety of processes that one might undertake. However, from experience, this often amounts to little more than gut feel.
…and that’s ok. Initially.
Be the campaign undertaken by an in-house team or through liaising with a client, typically there will be someone from the marketing team who has a solid grasp as to who they want to attract and why. Paid Social is great for testing out such hypotheses, delivering quick, clear learnings at different stages of the funnel (awareness/in-market/action etc. etc.).
The issue that I uncover relates to what happens next. Typically, the team responsible for implementing and managing the campaign will simply look to translate this group into one that fits the relevant platform and away they go. Ads are uploaded, the campaign goes live and, hey, it may even deliver some great results that the business/client is happy with.
But with segmentation, we can do so much more.
An extra step should be considered. Within this macro-target, there will be smaller, more nuanced interests, profiles and demographics. Try building these out into separate groups. Quite quickly, we can build out a simple structure like this:
To those familiar, this looks a lot like a structure for a PPC Paid Search campaign. For those who aren’t… trust me! The point is that this granular structure is the key to insight.
And insights are in turn the key to a great campaign but also understanding our audiences.
Very quickly, the advertiser can see which posts and, crucially, audiences are receiving the best engagement. With some simple tagging, we can then also see which are interacting with the landing pages on our website, be that something absolute like a sale or a sign-up or softer KPIs like dwell time or page views.
We can understand which audience subsets are showing the most interest in our business or products. This is profound.
Firstly, we can now direct our campaign towards those best-performing audiences, spending our budget in the most efficient way. Great!
But beyond this, these media insights provide us with real, empirical data on our content, our campaign and our brand.
So why doesn’t Facebook encourage this?
As market leader in this field, there are plenty reason enough to be using their products as the cornerstone of any Paid Social campaigns. The methodology I have outline in this article works perfectly when applied to Facebook/Instagram advertising. But this is not recommended as Best Practice in their (relatively recent) Power 5 recommendations.
Power 5 is Facebook’s own playbook for advertisers looking to “unlock new phases for growth”. Indeed…
“The days of manually hacking your way to ad success are no more.”
In other words, the opposite approach what I have been advocating above!
Here’s why I cannot agree with their recommendations:
Each of their recommendations is essentially asking the advertiser to automate elements of their campaign, to rescind control and put their trust in Facebook’s algorithmic black-boxes. Their argument is that it saves us time and delivers better media results. So, if you trust shoveling your ad dollars into something like this, be my guest. But there’s more…
I want to focus particularly on the Account Simplification pillar, for this relates specifically to the theme of this article. Facebook here is asking you to ignore the kind of nuanced structure that I advocate above, instead actively encouraging the advertiser to clump all their targets into bigger, macro targets. Shovel everything into one big campaign/adset, they say. The machine learning algorithms are so smart that they are actively better than a human at finding the best people to target. Ad performance will increase. Again, if you trust this particular box, that’s fine too. But…
Whilst we may gain a few extra points in media KPIs (we may receive a few more impressions for our budget, a couple of more clicks), think about what we lose: that insight. There is no way to tell what’s resonating with whom other than within a large amorphous blob of a targeting group. The algorithm might no, but we as advertisers have no visibility of this. As such, if you want to take this data from Facebook and apply it to something else (say, which magazines to target with press advertising), I now have nothing to go on.
In conclusion:
Segmentation brings us information, and from this we can extract rich insights that help us understand our prospective customer base.
How to go about undertaking this approach? Get in touch and we can talk about how we do just this!