It stems from a combination of two things: a lack of trust in post-view data and the assumption there must be some kind of intent behind a click.
The first assumption has a reasonable basis in fact: when you look at the average viewability for the industry, you can’t blame advertisers for looking at post-view conversions with suspicion.
The second is where the problem arises – particularly in the world of fraud and accidental clicks: is a click actually a signal of intent?
Identifying the ‘high probability’ clickers
To answer this question, we analysed behavioural data to build a lookalike model to work out if people who look like clickers, do actually click more than usual.
Initially, this held true – targeting high probability clickers generated a high number of clicks.
However, KPIs tend to be related to conversions, rather than clicks, so we ran the model against converters.
Knowing that the model accurately predicted clickers, we expected that (based on the traditional assumption that clicks and conversions are correlated) high probability clickers would also be more likely to convert.
Do clickers convert?
We actually found that high probability clickers were 20% less likely to convert than average.
In fact, you could improve campaign performance by excluding high probability clickers.
So why is this and what does it mean? There are two main reasons:
1. Click Fraud
It’s not unusual to see clicks occurring on impressions that were never seen. This is something that requires action from both publishers and agencies to eradicate.
Click fraud is easy to spot, by distinguishing between viewed and non-viewed impressions. If you are seeing lots of clicks on a particular domain/publisher coming from non-viewed impressions, it’s almost certainly click fraud.
Think about blocking that domain from your campaigns or raise the issue with the publisher/SSP responsible, who should be able to help.
A good brand safety solution should be able to stop you from delivering against suspicious users or domains.
2. The Accidental Click
Typically seen on mobile, tablet or gaming sites, this is a real click, but made with no intent behind it.
Though not fraudulent, they are still fairly easy to spot. The most straightforward indicator is an unusually high click-through rate (CTR) coupled with very low conversion rates.
The simplest fix for this is to monitor CTRs by site, excluding those with very high CTRs and no conversions. This will have the effect of reducing your average CTR, but should also reduce your CPA.
You will see fewer visitors to your site but more actual sales.
What does this mean for you?
Well, every time you pay for a fraudulent or accidental click, you’re wasting money. Even if you’re not paying on a CPC basis, budget delivered on inventory with high levels of fraud or accidental clicking is wasted.
These users are not going to convert. They’re never going to buy your products. Even if they actually visit your site, it’s by mistake.
But there are users out there who are real and are interested in your products: those are the ones you should be directing budget at. But what can you do about it?
There are simple steps that you can take to reduce the occurrence of these. But they’re unlikely to happen until advertisers stop rewarding their agencies for clicks and click-based performance.
Even those agencies with the best of intentions are unlikely to scrutinise their clicks too closely when they’re faced with a low CPC goal for a campaign.
If you want to reduce the numbers of fraudulent or accidental clicks in your campaigns, you’ll have to work with your agency to make that happen.
Help incentivise them to seek out and cut out non-viewed or accidental clicks. Don’t penalise them when this causes your CPC to increase.
Though it may look more expensive, more of the traffic to your site is now real, reliable and ready to convert.
So what does lead to conversions?
You may now be asking yourself: if the clicks on my display advertising aren’t creating conversions, then where is the value?
The way display works is through its ability to generate ‘view-through’ conversions. These are conversions created on other channels (such as search) that happen because a user has seen multiple display ads.
You can attribute display adverts to view-through conversions because programmatic allows you to track the many touchpoints display has along the customer path to purchase.
This provides advertisers the chance to see the real effect it has on CPA and conversion rate.
Econsultancy has just launched a new conference, Creative Programmatic, which takes place in London on March 2.
Attendees will discover how brands and agencies can benefit from delivering a personalised and relevant creative programmatic experience at scale.