Why Should I Care About Attribution?
It’s not always easy to understand the impact all of your marketing channels have on your overall performance – with many of these working together to produce leads in a way that isn’t necessarily the most predictable.
Depending on the nature of your business, customers may make a large number of interactions with your ads and website before deciding to convert (see the multi-channel funnel report below for one of our clients):
So how can you make sure that you’re giving credit where it’s due?
Well, usually it requires a little bit of digging, but it is possible to get a decent idea of what’s important and what’s not, by ensuring that you see the bigger picture.
Essentially, you want to make sure you’re giving credit to the channels actually driving bottom-line impact to your business – and taking a hard look at how you are attributing goals and conversions is a good place to start. Read on to go deeper into the thrilling world of online attribution.
Let’s have a quick look at the most common attribution methods you’re likely to come across when looking into your business data:
- Last click
Pretty self explanatory, whichever method the user arrived on the site by immediately before converting will receive all the credit.
- Last non-direct click
Also quite obvious, the method that the user arrived on site by immediately before converting that wasn’t direct will receive the credit.
- First click
The first click that brought the user to the site within the relevant time period before converting receives all the credit.
Credit is split equally between each step in the conversion funnel.
- Time decay
Channels closer to the point of conversion receive the most credit.
- Position based
The first and last steps in the conversion journey each receive 40% of the credit, with the remaining 20% being allocated evenly among the middle steps
- Data driven
Uses Google’s algorithms to determine the key steps that led to the conversion and attributes credit accordingly.
- Any click (not technically an attribution method but relevant nonetheless)
The channel receives maximum credit no matter where in the funnel it was interacted with.
The default attribution method is often set to last non-direct click in a number of platforms such as Google Analytics and Google Ads.
This can often be often misleading – for example if a customer finds the company initially through a paid ad and becomes familiar with the brand, but then returns a little later through organic search to convert. Does it make sense to see this as a conversion generated by organic search? If they had not seen and clicked the paid ad in the first place, they may never have found the brand and visited the site at all.
On the other hand, if the search term leading to the organic click was generic rather than brand related, the presence of the organic ad was crucial in the conversion for reminding the user of the brand’s existence/offering.
First-click attribution would credit the paid ad, and other non-last click attribution variants would split credit between the two (such as linear, which would split it 50/50, or time-decay which would weigh the credit towards the closest touchpoints to the conversion). Tools such as the multi-channel funnel, that we will look at shortly, can help you to deep dive into each conversion path and work out a likely explanation for the conversion behaviour.
For some companies there is one stand-out attribution method that makes the most sense to focus on, however for a lot of others it may be useful to analyse a combination of methods to understand the value each separate channel is bringing in.
Understand Your Customers
The key to attributing correctly is to understand your customers; how long do they take to research? Are they signing up for a free trial and then returning to purchase at a later point? Do they shop around or is your product/service an impulse-buy? Is the need that your product fulfils time-critical, or might users come back months later to spend their money?
Between industries this varies massively, and once you have an indication of this for your own business you can start breaking down the funnel to work out which touchpoints were crucial to the conversion and which weren’t.
Google Analytics’ Reporting
There are a variety of areas within Google Analytics that will help you to dive deep into your customers’ conversion behaviour:
The Multi-Channel Funnel
The multi-channel funnel in Google Analytics is a key tool here, allowing you to see the path to conversion clearly by source/medium, campaign, keyword, and various other dimensions. An example of the multi-channel funnel ‘Top Conversion Paths’ section in GA is shown below (as well as at the top of this blog).
As you can see, it helpfully splits out the touchpoints in the conversion journey, and covers the time duration up until whenever you’ve set as the look-back window.
You can also delve deeper into the specific channels, for example – paid search can be refined down to a search term level, organic can be split into the specific source/medium, and you can look at the different landing pages at each step.
The time lag section within the Multi-Channel Funnels area of GA is also useful here. This shows the time in days between the first interaction with the site and the conversion (although sadly this is capped by the cookie duration). In the screenshot below we can see that while the majority of conversions occur within one day of the first user interaction with the site, some take between 31 and 60 days, and one even falls into the 61 to 90 day category.
Similarly, the Path Length section also sheds some insight.
These two can be combined using conversion segments in Analytics to see the correlation between the two to inform your strategy, and usually tends towards longer path lengths taking a higher number of days to convert – which makes sense as users who do more product research (and so visit the site more times) ahead of purchase will generally take a larger amount of time to do this. You may, however, find that this correlation varies depending on your particular product/service offering.
Most of the time when we are analysing the effectiveness of paid search in the media mix, we will look at first-click, last-click and any-click attribution. Last-click is the default attribution method used by GA, although you can create custom segments in Google Analytics to show first-click and any-click – screenshots of how to set these up are below.
These segments should include users where the Default Channel Grouping matches Paid Search, for ‘First user interaction’ and ‘Any user interaction’. This will then allow you to break out your paid search reporting by last, first, and any click attribution – and can be created for any of your digital channels by altering the Default Channel Grouping.
An example of this breakdown for some paid search activity can be seen below in a small excerpt from one of our famous reporting dashboards!
GDN Impression Reporting
This is a relatively new feature and is pretty cool – Display ads are the billboards of the digital world, and it’s tough to track whether users who have seen them then go on to make a purchase, or completely ignore them. Well at least, it used to be, until Google brought out their GDN Impression Reporting, allowing you to see Display impressions as part of the assisted conversion report.
This takes a bit of effort to set up (see instructions for implementing it here) but once it’s done you will start seeing the following icons popping up within your assisted conversion funnels, representing static and rich media display ads respectively:
No longer will you look at your Display campaigns’ relatively poor conversion performance with tears in your eyes, as this will allow you to see how your ads are subtly reminding customers who then later find their way back to your site and convert (hopefully!)
GA Attribution Beta
In another recent addition, Google has implemented a new beta into Google Analytics that allows you to create bespoke attribution projects.
The functionality available here is very similar to that already present in GA with nothing hugely revolutionary about it, however it has a significantly swankier interface and allows you to keep your attribution projects separate (and run multiple alongside each other with different lookback windows, properties, and conversion types.
Above is a model comparison table comparing last non-direct click with linear attribution. You can see the difference in attributed conversions in the right hand column, which is pretty neat.
Below is a conversion paths table which is basically the multi channel funnel report from GA’s regular interface, except with the addition of credit percentages for each of the channels. It doesn’t credit Direct at all, as each of the attribution methods specify that they are “non-direct”.
Google Ads’ Reporting
Google Ads allows us to get deeper insights into the role paid search plays in conversions within its own interface.
The model comparison tool is rather good – an equivalent exists within Google Analytics to track across all channels, but Google Ads’ specific paid search option can be useful too
Cross device attribution is also important here, and uses a users Google account to track them across their different device types. If a user clicks an ad on mobile and later converts using the same Google account on desktop, this will count as a cross device conversion.
You can find this report within the Google Ads interface under Tools, Attribution, and then Cross-Device Activity. This is important when analysing device conversion data, as just looking at top line results may not show the full impact of device types in the conversion journey.
Within Google Ads the default attribution method for conversions is last click, however you have the option to change how your conversions are attributed, on a per-conversion action basis.
Under measurement settings in your account, select ‘Conversions’ and edit the settings – there should be an ‘Attribution Model’ dropdown that allows you to change the attribution model! This is handy, especially if you are using automated bidding that relies upon accurate conversion tracking to influence bidding, as you can select the model that best represents the most important parts of your users’ conversion journeys.
As a side note, if you are uploading offline conversion data, be aware that you will have to deal with attribution externally as it is not able to work it out within the platform. More info on that can be found here. Also, it’s worth noting that their limitations to the way GA and Google Ads attribute – users moving from device to device, browser to browser etc. are not always able to be tracked throughout every stage and so won’t necessarily always be tied up into the same conversion journey.
The main takeaway here is that you have the tools available to really dive into the impact that each channel has on your conversions, and how channels interact with one another throughout the user’s journey. Using them, you won’t have to be in the dark about where’s best to put your advertising budget, and you won’t end up neglecting crucial channels.
For more help with attribution and reporting, get in touch with us – we’re always happy to hear from you.