Human vs machine. In PPC, the battle line is real, and it’s fierce.
Both the antagonists are flawed…
In one corner: Algorithms. Quick hands, but a bit tunnel-visioned. Not great at adapting to situations and the wider context…
(Yes Mr. Algorithm, we’re getting more ‘leads’ from 18-25 year olds, but they’re disproportionately job applicants…
No Mr, Algorithm… we’re hitting our budget, so right now we’d really rather lower bids to spread our current spend levels further… etc)
In the other corner: us. Good at the big picture, but we’re a bit slow, and prone to make suboptimal choices based on our biases and temperament.
Better Safe than Sorry?
One particular bias that often gets in our way is the precautionary principle, which holds us back from switching smoothly between restrained and ruthless as called for by the situation – and priority – in our campaigns.
To stay ahead of Smart Bidding, you have to exercise smarter bidding – and one way of doing that is to get better at overcoming this temperamental bias.
Too often, we tinker around the edges of optimisation, not wanting to ‘jump the data’… not wanting to lose traffic that’s providing – or might yet provide – some nugget of value…
But being afraid to close any doors can keep an account in a stasis of bloated traffic with unnecessarily high CPA… Sometimes it takes bold moves to really swing that traffic in the direction of better converting users.
Consider the following situation. Manual bidding. You need to improve CPA as a high priority, so you’re looking at any and every variable to eke out a bit of extra performance.
Here’s what you see in Demographics > Age Ranges. 130 days’ data.
First impressions?
“18-24 isn’t great… nor is ‘Unknown’.
Let’s slap a bid reduction on Unknowns, nudge 18-24 even lower (let’s see how they like -30%!)
…and as for 65+, we’ll probably want to remove (maybe reverse) that +20% bid adjustment.”
If that’s how you’re thinking, then you’re making sensible moves, based on solid-enough evidence. But the thing is… how much is any of this going to move the needle?
Shifting bids by 10-20%, on ~30% of your traffic is not going to give you the kind of impact you need when you’re on amber-alert for CPA (and if you’re using Enhanced CPC, remember your bids are already varying by a lot more than this, outside of your control.)
Activate Beast Mode
Instead, think about excluding underperforming segments first.
Ask, “Can we afford to lose this chunk of traffic?”
Often you can afford to lose that traffic perfectly well – more often than we give credit for… (partly because of another of our natural biases: the tendency to overvalue the status quo… see Kahneman and the Endowment Effect).
And if you can afford to lose it… chop it.
And if you’re hitting your overall budget… then lopping off a limb from your account may not even reduce your overall spend.
So should we just remove those under-performing age ranges?
Probably.
As it turns out, in this case taking a shorter date range showed that Unknown has performed pretty well in the shorter term, so it gets a reprieve.
But 18-24 and 65+ – you’re walking the plank.
And that’s more like it.
Almost 20% of overall spend that was going towards underperforming traffic is now gone.
Either that spend gets directed straight back into higher-performing segments – or it doesn’t. Either way, swinging the axe is definitely going to move the needle on CPA.
One precaution that is worth taking… If you are judging by CPA, you will also want to keep an eye on the ‘upper funnel’ influence of any chunk you remove. Be aware of traffic that has disproportionate hidden value in attracting those users who later convert via another campaign, or another channel.
This is particularly likely for mobile traffic, which often shows poor conversion rate on the face of it, but turns out to be genuinely feeding the funnel.
Here are 3 ways you can take that precaution in practice:
- Check for assisted conversions in the Top Conversion Paths report in GA
- Use a better attribution model (better being not last-click)
- Check the Search Attribution reports
But with that nod to health and safety out of the way… keep in mind that our natural caution can hold us back from making the gear shift we need when our campaigns need a serious performance boost.
When improving CPA becomes high-priority, sometimes you need to go full Marie Kondo, and be willing to chuck a few babies out along with a bucket load of bathwater.
So be ready to dial up your aggression level when your account needs it… Learn to activate PPC Beast Mode 
.
