How to improve AdWords performance through segmentation
Data segmentation is absolutely key to the performance of your AdWords campaigns. In fact, to take a quote from Google Analytics expert Avinash Kaushik, “segmentation: do or die”
In Google AdWords, aggregate data refers to statistics that have been combined from several measurements. This condensed information is useful for analysing summary performance and trends.
However, it would be ineffective to make changes in AdWords based on aggregate data, because the same bid wouldn’t be as effective for – for example – desktops and mobile, weekdays and weekends, London and Cheshire. This is where data segmentation comes into play.
Data segmentation functionality is built right into the AdWords interface. AdWords bid modifiers allow us to make changes based on segmented data by applying modifications to bids based on the time of day, day of the week, device and location.
In segmenting your Google AdWords data, you can drill down past the aggregate summaries to view deeper insights.
Example data segments in AdWords
When looking at each of these segments, we should section the data again by campaign level. Each campaign may behave very differently and each campaign should have its own KPI targets. For example, the level of ROI we expect from a direct response brand campaign may well be very different from a prospecting generic campaign.
We could extend this methodology and segment down to AdGroup level. However, for some segments, AdWords only allows us to make changes at the aggregated campaign level, such as when making changes to search partners. If segment-based changes can be made at AdGroup level and there is the volume of data for statistically significant insights, I would recommend segmenting down this far.
Segmented by day of the week
In the example above, I segmented our example AdWords revenue data by day of the week and calculated the ROI ratio for each day. I have presented the data so we can see which days tend to perform well and which days need attention. We can clearly see that Monday is the best performing day for this campaign. As a result, we should increase bids on this day to capture more of this high ROI traffic. As Saturday is the worst performing day, we may want to lower the bids or even pause our advertising on Saturdays.
Note – I would look into Saturday’s attribution towards the other days of the week before making any changes, as visitors may be doing research during the weekend and converting on other days of the week.
Performance by hour of the day
In this example, we can clearly see that from the hours of 10 to 12 and 13 to 14 the campaign has a strong ROI. We will increase our bids for these hours. We can also see that in the evening, the campaign performance is poor. We may want to lower the bids in these hours so more of our budget can be spent in high ROI hours.
Performance by device
We can see that tablets provide the highest ROI and computers produce the bulk of the revenue for the campaign. Unfortunately bid multipliers in AdWords for tablets and computers are grouped together. However we will look at reducing our bids for mobile traffic.
Note – Having specific mobile ad copy and mobile optimised websites will improve the performance of mobile traffic.
Performance by search network
The performance of traffic from search partners is significantly worse than that from Google search. Our only option in AdWords in this instance is to disable the search partners for this campaign and move the budget to Google.
Performance by location
In this campaign we have 107 different location modifiers. We can see from the visualisation above that some locations perform better than others. Cambridge has a relatively high volume and a good ROI. We had made bid adjustments based on location performance. We will also add extra targeting around high and low performing areas so we can extend bid modifiers.
Using segmentation we have been able to provide actionable insights that will have a positive effect on our bottom-line. We can now make changes to our AdWords campaigns and measure their effectiveness.