Estimated reading time: 9 minutes
Download a print-friendly PDF
The easiest way to get started with remarketing – in Google Ads at least – is to use the native remarketing options within the Google Ads interface. Within Google Ads you’ll find this under the ‘tools’ menu at the top, and under ‘shared library’ and ‘audiences’.
You can natively build audiences of people who have visited your website recently, (or visited certain pages), or bought something recently, or subscribed recently, or watched one of your YouTube videos recently. With custom programming you can also build audiences based on custom events – things that happen during a website visit.
Plus you can set audience durations – the length of time somebody remains in a particular audience. This is critical in setting how long your retargeting ads appear for. Audience duration is probably the most important aspect of any audience (i.e. how recent was the last website visit), followed by the content – what pages on the site did they view. In 80/20 thinking, how recently somebody visited your site followed by the pages they viewed are the two most critical factors.
All of this is available natively in the Google Ads platform.
However, you can also also build retargeting audiences using Google Analytics, by linking your Google Ads and Google Analytics accounts. This gives you additional controls for refining your retargeting strategy. With Google Analytics you can filter your audiences by behaviour – what people did while they were on the site, not just what pages they viewed.
Google Analytics has more data about how people behave while on your website than Google Ads does.
In Google Analytics go to the admin area, go to property settings in the middle column, and down at the bottom there’s an option called ‘audience definitions’. This is where you create your audiences.
Segment Audiences by Traffic Source
A nice option is to build audiences based on a website visitor’s UTM tags (campaign / source / medium / term / content). Source and medium are most common. For instance, a person who clicks on one of your Facebook ads came from Facebook (so ‘Facebook’ is the source) and ‘cpc’ (cost per click) is the medium. You can then build a retargeting audience of only people who came from Facebook, or only people who engaged with a particular campaign, or even clicked on a particular ad.
With the exception of Google Ads, this ‘UTM’ code usually has to be manually appended to your links. So the end of a link might look like:
In this case Connectably is the name of our email system (and hence where the click came from), the medium is email, and the campaign is ‘Story Copywriters’. Using Google Analytics we can add people who click on this link to a separate retargeting audience, without sending them to a unique landing page.
Using email as an example, you might build audiences of people who:
- Clicked on any email
- Clicked on an email in a particular email series (perhaps promoting a particular product)
- Clicked on a particular link in an email
You might for instance find that people who click on the bottom link in an email are more likely to convert than those who click at the top (arguably because they’ve actually read the email!), so you might want to remarket to them differently.
Running retargeting ads to everyone who visited a particular page in a particular timeframe is fine as a starting point, but it misses the context of where those individual visitors came from. A visitor who came from email will arrive with a different set of expectations to one who arrives from Facebook. People who arrive from one Facebook campaign may arrive with a different set of expectations to a different Facebook campaign. Consequently you may want to follow-up with them differently.
This becomes especially important if you are buying a significant volume of clicks offline, e.g. radio, TV, billboard ads etc. In that case you might setup a dedicated domain for the ad (often called a ‘vanity domain’), and build an audience of people who respond through each channel. The domain itself may just forward to a particular page on your website, with the UTM tracking information appended to the page URL. If you have a lot of spare domains in your hosting account, this is a good use for them!
To take this a level deeper, if somebody clicks on an ad in a particular Facebook campaign, you can then start to do sequencing. In other words, the next Google remarketing ad they see might follow on from the ad they clicked on in Facebook. You’re adding a level of contextual intelligence to your advertising. You know they’ve clicked on ‘ad 1’, so now you show them ‘ad 2’. Ad 2 continues the conversation started by Ad 1.
You can do this right across your retargeting maze. You have an entry point somewhere, perhaps a Facebook ad, or Google ad. Once people enter at that point, you follow up with a different ad. You don’t just hammer your audience with the same message over and over.
Exclude People Least Likely to Buy
Almost everyone who uses Google Analytics is familiar with a metric called ‘bounce rate’. A bounce happenes when somebody visits your website, and leaves without viewing more than one page.
If you’re running any kind of ‘all visitors’ audience, some percentage of that audience will bounce. The people who bounce are usually the least likely to buy or convert in the future, meaning you may wish to avoid retargeting them. Google Analytics allows you to do this.
In considering bounce rate you also have to consider time on site. Somebody who arrives, spends 30 minutes on one page watching a video and then leaves, still counts as a bounce. But they clearly engaged with your content based on the amount of time them spend on site.
By showing your retargeting ads to people who bounced / spent little time on site, you’re likely showing ads to people who simply weren’t interested. This lowers the click through rate of your retargeting ads, raising the amount you’re likely to pay for each click. If your ads have a low click through rate the network makes less money when they run your ads, and eventually you will be penalised for that.
A great way to raise your click through rate is to avoid showing your ads to the people who appear to be disinterested based on their past website behaviour. To exclude the worst of the ‘bouncers’ we suggest excluding people who spend less than 30 or 60 seconds on your website. Those people clearly decided they were in the wrong place to find whatever it was they were looking for.
If your website is content-heavy, a variation of this would be do segment your retargetign audiences based on ‘page depth’ - the number of pages a visitor views. You might for instance want to spend more of your remarketing dollars on people who view at least 3 pages on your website. That behaviour indicates an active interest or information hunt. Maybe the reason they didn’t convert is they didn’t find the information they were looking for. In your remarketing ad you can then offer that information one click away, on a silver platter. Assuming you understand what they might have been hunting for in the first place. Which is a good reason to use live chat tools, search boxes and feedback mechanisms on your website (we recommend hotjar).
If you have a lot of website traffic, adding this behavioural information to your remarketing strategy is virtually guaranteed to save you money and improve your click through rate. The people who do click are subsequently more likely to purchase.
Using Google Analytics you can slice and dice your remarketing audiences using a huge range of metrics, for instance demographics, user location, device, browser. You can build audiences of people with particular interests (these are the same as the ‘in-market’ or ‘affinity’ audiences in a Google Display Network campaign).
You can build an audience of people who visit your website on a mobile device, with the intent of running a remarketing ad to them on desktop later on. Often people won’t complete a purchase while they are on their mobile device, because they are on a bus, it’s inconvenient, they’re at work, etc. But they might convert later on. We also suggest treating tablet traffic separately – tablet traffic often behaves like mobile (and often converts worse), even though it’s a bigger screen. This may be because people use tablets for quick informational searches, while watching TV for instance.
You can build an audiences of people who visit your website on an Apple device. These people may convert differently to Android users. (Test this – please don’t make assumptions!) If you sell a technology product, you may wish to advertise differently to people who arrive on your website using the latest iPhone. (Those people are probably early adopters, happy to invest in new technology).
RFM (Recency, Frequency, Money)
We’ve talked on this podcast previously about the ‘RFM’ principle (recency, frequency, money). The principle states that the most likely person to buy is the person who visited most recently, followed by the person who visits most frequently, followed by the person who has spent the most money in the past.
By using Google Analytics ecommerce tracking (you can also implement this on regular order forms), you cab build audiences of your highest value, most frequent and most recent purchases.
You might for instance build an audience of high value but lapsed customers. Customers who spent a lot with you in the past, but not within the last 6 or 12 months.
If you know that your best customers buy every month, you might build an audience of people who have placed 12 or more orders in the last year. A primary objective of your marketing strategy should be to keep these people happy and buying. These people should see ads from you more often. Advertising to them is likely to be a high ROAS (Return On Ad Spend) strategy. The more online transactions you process each month, the more important this becomes.
It’s perfectly fine to get started with remarketing with the native Google Ads options. Google Analytics allows you to refine your strategy and segment your audiences at a deeper level of optimisation in your sales process. You’re weeding out the people statistically least likely to buy.
If you’re spending thousands a day on ads you should consider hiring or contracting someone to configure these audiences, because the pay-off can be huge. A 1% increase in your results could easily pay for that, so do the cost-benefit analysis.
Google Analytics remarketing does not replace the need to configure audiences in the native Google Ads audience manager. Within Google Ads you can upload your customer or prospect list, and you can create lists of people who watch your YouTube videos. This cannot be done effectively in Google Analytics. So use Google Analytics to enhance your Google Ads remarketing strategy, rather than replace it as a technology.