In your Google ads SEA campaigns, you have all your keywords that you are bidding on. Depending on the match type you give them, your ad may appear when there is a match with the user’s search terms he is typing. In Google Ads you have 3 types of keywords: “exact match”, “phrase match” & “broad match”. In this article we will focus on the latter as its way of functioning has been improved by Google algorithms. You will get the practical guidelines to be able to implement your own A/B test on Broad Match.
Before starting, you must note that Broad match is the default match type that is assigned to your keywords. On top of that, the following definition only applies to positive keyword targeting (Negative broad match works in a different way):
Broad Match: “Ads may show on searches that are related to your keyword, which can include searches that don’t contain the keyword terms.” (Google Definition).
In other words, your ad will appear on the searches that are relevant to your product/service you are selling. In order to do so, Google has improved the broad match type by taking into account other signals such as the landing page, the user location and the other keywords you are bidding on in the same ad group.
You might be more used to the “phrase match” type as you feel more in control on which searches your ad will appear. However, you might lose opportunities for some searches. Indeed, searches on the search engine are always unique as the signals sent by the users differ from each other. On top of that, new queries are popping up every day.
By using broad match in your account, you will benefit from:
- Efficiency
- Time saving
- Reach & coverage expanding
- Relevance
- Increase your campaign optimisation score
If you are still reluctant to convert your keywords in broad match, you have the right to be as this might not be convenient for all sectors and campaign types. Indeed, we would not recommend it for your brand campaigns as you want your search terms to always include your brand name. Nevertheless, it might be interesting to test it for your generic campaigns.
Upfront launching your experiment, it is fundamental to align on how you are going to assess the success of it. What are the KPI’s you are going to select and as of which percentage uplift you consider the test a victory.
In order to make an A/B test, you can follow these steps:
- Make sure you don’t have any broad match type keywords in your original campaign.
- Remove your campaign from the shared budget if you are using one.
- Go in the recommendation tab of your campaign and look for the “Upgrade your existing keywords to broad match” card.
4. Click on “View recommendation” then on the three vertical dots in this section:
5. Click on “Apply as an experiment”:
You will now arrive in the classic setting of an experiment where you can put your start and if needed, end date. For an ideal A/B test, we recommend having it running for 8 weeks, especially if your conversions have an important delay. In that way, you make sure the algorithms gathers enough data to be able to measure a significant uplift.
Meanwhile your experiment is running, it is important to continuously optimise your campaign by looking at the search terms triggered by your broad keywords. This will help the algorithm understand your business better.
After a couple of days, you can go to the experiment section of Google ads to have a look at the experiment summary.
Depending on your KPI’s, evaluate if the testing campaign is having better performances. When you estimate the experiment has enough data to measure an uplift, you can decide to “apply” or “end” the experiment. When applying, the changes will automatically be done to the original where all keywords will be converted into broad match.
For some clients, we had the opportunity to make the A/B test on few campaigns. One of them ran for 6 weeks and as the trial generated a higher ROAS (+15%) and higher conversion value (+17%) with still relevant search terms, we decided to apply it to the base campaign.
However, for another campaign where we did the same test, results were not convincing. Hence, we ended the experiment.
You now have the steps & tools to set up your own A/B test on broad match keywords. As you can see, the test is easy to implement and it is up to you to put it in place for each of your campaign’s type to determine whether broad match brings better performances or not.
Elisabeth de Biolley
| LinkedinThis email address is being protected from spambots. You need JavaScript enabled to view it.
