Last week I asked what you most wanted to hear about. One of the replies was: “To understand the ROI on adwords better.”
“Hahaha,” I thought in my head, “this could take a while…”
To give you a complete answer would take more words than you possibly have the inclination to read right now. But the bare bones are as follows.
The key idea is that tracking ROI on ad spend is best understood in relative terms, not absolute. It’s often difficult to pinpoint where revenue has come from, because somebody who buys may do multiple searches and click on multiple ads in the build up to a sale. Or may click on both Google and Facebook ads, to really screw things up.
This is more common than you might think, especially if you’re using retargeting.
To make matters worse, different tools report on conversions in different ways (this is called ‘attribution’). AdWords has historically used last click attribution, where the credit for the sale is given fully to the last keyword the customer searched for. Google Analytics lets you switch between attribution models.
Both measure conversions and results in slightly different ways. The key thing is that your AdWords data and your Analytics data will never fully match up.
And honestly, it doesn’t matter. I’ve had AdWords clients in the past who drove both them and me to distraction by trying to figure out exactly where every single dollar came from in giant spreadsheets. (I hate giant spreadsheets).
If you know that your profit from Ad A was around $30, and your profit from Ad B was around $70 does it matter if the exact numbers are slightly out? It’s enough to make a decision. I’d actually argue that if both ads are profitable, you should probably still run both because different people will respond to different ads.
For decision making purposes I’d calculate profit as gross rather than net (i.e. revenue minus spend). You could try to take off your product delivery costs, but most of the time on AdWords you’re buying customers, not individual sales. You just have to ask whether the cost to acquire a customer is acceptable.
If you really want to measure ROI, you need to be tracking click information in your Customer Relationship Management system (join me on Monday’s webinar for more on this). If possible I would also store the last referring traffic source (campaign / keyword / ad creative etc) on the contact record in your CRM system, so you can measure which campaigns drive the most long-term revenue, rather than the most initial sales. This becomes more relevant if you nurture potential customers over a number of years.
Even then, the only number you can really trust is money that passes through your merchant account. Everything else is best used as a relative indicator of performance.
If you’re spending good money on Google (or plan to), and what I’ve said above has made your brain hurt, we probably need to talk.