The second part of our series looking at how advertisers can adapt to a more automated paid media landscape focuses on measurement. If you missed the first part of the series on structures you can catch up on that here.
With bidding being automated, measurement will become of fundamental importance. After all, a bid strategy is only ever going to be as good as that data that you put into it.
Bad measurement > poor data > poor bidding decisions > poor performance
Benchmark Your Measurement
Import Offline Events into Google Ads and GA so you can bid on them
When it comes to tracking user behaviour offline, most financial service brands that we speak with usually use UTMs stored in their CRMs to work out what actions users have taken.
The problem with this is that the data is still offline and Google can’t use this data to improve lead quality. To improve performance advertisers need to push this data back into Google Ads and Google Analytics.
They will first need to map out their customer journey to work out which events they want to track.
Here is a user’s journey for a lender where a user submits an application online before it moves offline to credit, before being either accepted or rejected. The lender here would want to track each one of these events along with its respective value.
Importing the CRM data to show where applications are in your lending process back to Google Ads and Analytics is done via the offline tracking functionality within Google Ads and GA.
Offline tracking works by storing a unique identifier when somebody makes a lending application either online or on the phone in your CRM. In the case of Google Ads it’s called a Google Click ID and user ID for GA.
Every time someone clicks on your advert Google assigns a unique Google Click ID that is associated with the keywords and campaigns that triggered your ad to show.
This identifier is placed at the end of your URL as a UTM parameter and looks like this: Demandmore.co.uk/?GCID=354325n3425 34j53252345
As somebody fills out your lending application with their information you will need to add a hidden form field for the Google Click ID. It acts like any other form field like name or email address in your lending application but it will not be visible to users on the website.
Finally, once somebody’s application gets approved you can import the Google Click ID back into Google Ads with the value of the closed application.
Google then assigns a conversion to the keyword that generated that application in the conversion column as well as the value of the application in the conversion value column.
You can do the same in Google Analytics by storing somebody’s user ID (similar to Google Click ID) and then importing it back into Analytics using the Measurement Protocol API.
Value-based bidding with Search Ads 360
Use floodlights to bid based on your offline data across channels
Advertisers using offline tracking to import data can take this a step further by using the Floodlight tags within Search Ads 360 (SA360). This allows you to import data into SA360 from your CRM such as Salesforce to provide a full picture of the customer journey.
One of the major advantages of using SA360 is you can share data across your other advertising to make bidding decisions in platforms such as Bing, Baidu and programmatic ads bought through Display and Video 360 (DV360).
Importing the data also allows you to import margin data back from your CRM such as the commission you’re paid on a loan if you’re a loans broker for example and attribute this right back to the keyword level.
You can then use this data to structure your accounts around the margin and projected return as opposed to focusing just on CPA such as cost per loan application.
Integrating LTV from your offline systems
Another major advantage of using SA360 if you’re a finance brand is it allows you to incorporate LTV data and optimise towards it.
Take the insurance sector, for example, not all car insurance enquiries are equal as some drivers will have a clear record while others may have a history of driving penalties and accidents. Similarly, with lending some borrowers will have a clean credit record and others will have defaults and CCJs.
To bid towards LTV you must first determine which factors influence the lifetime value of your customer so you can start to optimise towards them.
In business-to-business lending for example the number of years the business looking to borrow has been trading is likely to be a factor that correlates with LTV. There are substantial numbers in car insurance for example, such as; years no claims, model of car, type of car, the area you live in a many more.
Once you’ve identified these factors you can import your customer data from your data warehouse to SA360 on a real-time basis. This will allow you to track sales volumes as well as give visibility on the LTV of each of your sales inside the Google Ads interface at the keyword level, allowing you to make advanced optimization decisions.
Adopt Privacy First Tracking Methods
Move to Data-Driven Attribution
Data-Driven Attribution and Enhanced Conversion Matching will bridge the gap in a world without 3rd party cookies
In late September Google announced that it would be making DDA (Data-Driven Attribution) the default attribution model within accounts and removing the data requirements. Data-Driven Attribution assigns fractional values to conversions based on how they influence conversions.
Data Driven Attribution uses “survival analysis” to determine the counterfractional gains of each touchpoint by training on data from control groups. With some users being exposed to certain ads compared to the propensity of conversion of users who did not.
Here is a very primitive example of how survival analysis works to power data-driven attribution.
Users are broken into two control groups. On sees the Youtube video advert while the second set do not. You can see by showing users the Youtube video that it increases the conversion rate by 25%. Of the three touchpoints, you would therefore weight the Youtube video touchpoint with 0.25 of the conversion and the 0.75 splits between the two search touchpoints.
Data-Driven Attribution will allow advertisers to model data effectively bridging gaps where certain sessions are not observable while respecting user privacy.
For example, if a segment of conversions can be observed in Google Chrome but not in Safari, Google’s machine learning will be able to model the probability of conversion in the unobservable browser incorporating other dimensions such as time of day, location, device etc to model the likelihood of a conversion taking place.
Enable Enhanced Conversion Matching
Enhance Conversion Matching is currently in beta but is a setting that you will want to enable to help improve your conversion tracking accuracy as we move into a world without third party cookies.
As advertisers are well aware, third party cookies will be being phased out in 2023 in a bid to improve user privacy which will make it harder for advertisers to track user behaviour.
Enhanced Conversion Matching allows for improved measurement as it works by sending hashed first party data back to Google which Google then matches to their users to record conversions.
For example, if a user goes onto your website. Enquires about a loan and enters their email address and phone number as part of the application.
This will then be hashed and then sent securely to Google. Google then matches this data against the data that they hold for users and conversion is registered in your account against the ads and keywords that triggered it.
Measurement will become of critical importance as bid strategies are only as good as the data that you put into them.
To thrive in 2022 Financial services marketers need to feed Google data from the full customer journey in real-time. Enterprise financial services brands with large budgets and complex bidding requirements around LTV should consider investing in SA360 to allow them to take these factors into consideration in their bidding decisions.
Finally, as the (third-party) cookie crumbles they will have to embrace data-driven attribution models and enhance conversion matching to start to bridge the data gap as cookies are phased out.