How financial services brands can accelerate performance with data  

Our ability to measure user activity across the web has massively changed and is limited by regulatory and technological change. Things like ICP and the GDPR make it significantly harder to identify individuals, their online behaviour and therefore the distinct touchpoints in their website journeys. What we are observing is an evident shift from a precision marketing paradigm to predictive marketing one. While we previously had a transparent view of user activity, we now only see fragmented bits and pieces, requiring us to make accurate  and informed predictions of what is likely to happen next based on the sample we see. 

The era of predictive marketing 

You gain key competitive advantage if you are able to correctly leverage existing predictive models and more importantly, build your own advanced ones. It puts you in a better position than your competitors who may not have that functionality or rely on stock models already built into your tools. This really draws a strong focus on initiatives that will increase the propensity to purchase or likelihood to churn. Reliable data is consequently a crucial component in order to do this. Without it, you’re likely to have unreliable models leading you to the wrong direction and decisions.  

The type of data we can draw on.. 

Complete and accurate data can be collected in several ways. We are all quite familiar with first party data. It’s the data you collect yourself, whether from digital analytics platforms such as Google Analytics or the information you gather about your customers and leads in your CRM. 

The next sphere is second party data, which people collect on your behalf. This crops up quite often and could be used regularly within your marketing efforts. Tools like Search Console and SEO monitor all track data about you or your website. It is therefore collected on your behalf and it is very important.

Third party data is collected without any prior agreement between yourself and the data processor. It is relevant and useful data you might want to feed into models to understand its impact on current or future performance. These three areas are vital to work out what you have access to and its level of reliability to help you understand remaining components you will need to fill in gaps as you go through.

Measurement: Why measurement matters? 

To ensure the collection of complete and reliable data, good measurement plays an extremely vital role. While measurement had very much been about granularity, which is what previously drove good performance, there has been a huge shift toward automation with a change in platforms over time, As seen with Google Ads, which was very keyword focused five years ago, now depends significantly on putting in the best possible data to gain competitive advantages in addition to the usual query level bidding. It could be feeding in good measurement data, audience data, business data and things like lifetime value.  So the advertisers that are doing the best now are the ones that got the most sophisticated sort of measurements using consolidated structures, great creative and audience structures.  

It matters because if you add poor data into Google’s bidding algorithm or don’t give it complete data, you are not getting into the nitty gritties. Google therefore isn’t going to understand the impact of the bidding, leading it to make poor bid decisions and overall result in poor performance. On the contrary, if you have great quality data, you’ve got your full journey mapped out, giving Google reliable data to yield really good performance. These will also be one of the things that will distinguish good advertisers on the platform from the really great ones; the ability to provide the bidding algorithms with the best possible data in line with key business objectives. 

Measurement: Bidding towards Value 

Similar to the data maturity framework for Google Analytics, we have another one more focused  on Google Ads and the process to bid towards value as opposed to lead quality. Here we go from beginner to pioneer as we call them. Beginners being people who are probably measuring online conversions and might have core conversion tracking. They are advertisers who haven’t yet invested heavily in data, but have considered it as part of their roadmap. The more intermediate advertisers would already be bringing in offline data. Here you’d expect to start a bid towards value. We would actually now begin to give Google a much fuller picture of what’s going on. 

The advanced advertisers would further pull in things like offline core conversion tracking, so linking their call centres to Google analytics. This way you’ve got your CRM call centre, analytics and Google Ads all connected together allowing you to gain a complete picture of what’s going on. You could go a step ahead and take into account different channels within GA rather than a single channel where you’ve are cookied by Google when bidding towards value. 

These are who we refer to as Pioneers. They would have in place full multi channel attribution and start really thinking about micro conversions. In insurance, for example, you’d also want to be looking at LTV as part of setting your bids and targets. You might want to take advantage of AutoML inside Google Cloud platform so you could push in your current customer data and attributes and provide a predictive LTV out the other end which can then be used to inform the sort of CPAs that you might want to bid towards. So that’s the stages we sort of see people going through and obviously the further down you are, the better results we expect to see from your performance activity. 

Measurement: Plot all of your Conversion (Lending example)

Implementing the framework can be challenging. To go into a bit of a case example,, we’ve chosen Lending. Here we focus on the numerous  customer journeys you see within financial services and in insurance. It is also analogous to things  like asset management and brokerages where you open accounts and other offline events where people make an investment 90 days later within the platform. We suggest kicking off by simply mapping out all potential customer journeys, which is something we rarely see people do. Whether offline through a call centre or online through a button click into an application, there are different stages your customer might go through to get a single or multiple quotes for loans.

This is quite a good example because it encompasses the distinct type of journeys in financial services’ business. It eases the process of laying out an extremely clear roadmap of all the different touchpoints, online as well as offline journeys and most importantly micro-conversions involved. In terms of measuring micro conversions, this is quite simple to track in GA4, allowing you to give the platforms much more customer data.

Measurement: Micro Conversions 

You can measure conversions, but if you have campaigns that are quite low volume, you might not be able to necessarily bid on funded loans if you’re only getting 30 a month.However, if you are getting 200 people to stage 3, it would give the platform plenty of data to optimise towards. So not only would it provide the platform the full customer journey but also useful things to bid towards if you are lacking in data, with regards to conversion actions slightly further down the funnel. You also attribute value to every micro conversion based on how likely users are likely to convert at that stage..

Measurement: Micro Conversions (Floodlights)

Below is an instance of doing it in a more advanced manner. Based on the way people fill out fields and forms, you can attach a micro conversion value if you’re using floodlights in a campaign manager. So if they’re a homeowner and you’re doing secured loans, you might want to up weight the conversion value if they select yes or down weight it if they select no. This would be a good way to indicate to Google the value of the leads if they are in the earlier stages of the purchase journey and haven’t got huge amounts of data. It’s a good way to give Google an indication of the quality of the lead.

Measurement: Offline conversions 

Offline conversions are primarily measured by using click IDs and the measurement API. When we typically go into clients, there’s usually a variety of ways they have done this. Some are using UTMs to measure offline activity and then stitch the two together with spreadsheets, which is quite common. Some might be measuring it just on Google Ads while others have it in Analytics, but ideally you’d have this data sent back to all the platforms that you’re using. So whatever that be, Facebook, Google Ads, Analytics, using the various different APIs. It is key you map out all the offline conversions throughout.

In terms of how offline tracking operates, it could be quite confusing but in a more straightforward approach, it often works by storing Google Click IDs or client IDs, that serves as identifiers that are unique to every ad and keyword whenever there’s a click. Google creates a string and attaches it to the end of a URL. As a result, upon a click on your advert, the customer is directed to your site and has a click ID attached to the end of the site URL. The ID is unique to the click and Google knows which ads and keywords triggered it.  All you then do is create a hidden form field inside your form. Here you obviously see the click ID that users wouldn’t be able to see but it will exist.

As they click submit, the click ID will go into your CRM system, additionally being used as an identifier throughout their journey and attributing to a conversion by sending it in a spreadsheet via Google Ads or the API if you’ve got the sort of technical ability in-house to do so. This allows you to have full visibility of the sort of offline activity that’s happening.  Several CRM systems now like Salesforce and HubSpot, Zoho have integrations into Google Ads immediately. You can map out your offline actions and then send them back. It is certainly possible to do with custom CRMs. We’ve built applications on top with clients to send back click ID when certain things happen and register value. So you can see the value of the activity. 

Measurement offline Call: Conversion tracking 

The same thing happens with offline core conversion tracking to get value back into the platform.This is something people don’t do very often. It’s really valuable, especially if you’ve got a business that has a lot of inquiries over the phone. It’s quite big in lending and insurance. Although instead of having the click ID here, the phone number will be used as an identifier. You use a Google forwarding number on your website.

When someone goes into the website, the number changes and then Google records the phone number that people have called you on. You store that phone number in your CRM system and basically report it back the same way.  Using a spreadsheet or the API, you could push back the phone number along with the conversion that happened and Google would register it as a conversion next to the ad. You could also pass back value once the loan has been funded, leaving you with a complete view of offline tracking of both, your clicks and call activity.   

Measurement : Measuring and Modeling LTV with GCP 

Within 90 days you can directly pass additional LTV back into Google using the click ID. If you have a second loan, you just send back an event using the click ID, which would be attributed to Google Ads. We could also do the same thing with analytics. On the other hand, following the 90 days window, we can’t use a click ID to import back into Google Ads, and this is where offline modelling makes more sense. 

In case of a Cloud platform , if you’ve got a CRM like HubSpot, Salesforce, you are able to easily push data for your opportunities into Google Cloud in a tabular format and consequently stitch those together, to get a glimpse of views per customer and an overall count of your customers, their loans, the timeline of the loans and their LTV. These points are then vital to determine the value of the customer and perhaps what your CPA bid will be if you’re focusing on bidding towards lifetime value as opposed to short term bidding where you’re looking at value or CPA in the short term.

If you have really advanced predictive modelling, Google has a range of built-in tools inside Cloud platform like AutoML allowing us to push this data into it and using pre-built tools that Google has got. It will then look at the attributes of a customer, your current customers and make a prediction on how valuable that customer will be to you. You can then use this to inform how much you want to bid on the CPA model if you’re looking at the overall lifetime value.

In conclusion 

To thrive in this landscape, businesses must prioritise building their own predictive models, leveraging existing ones, and ensuring the collection of complete and reliable data. By embracing the power of predictive marketing and investing in advanced measurement techniques, businesses can make informed decisions, optimise their bidding strategies, and unlock the full potential of their marketing efforts. 

It’s time to embrace the future of marketing and harness the power of predictive insights to drive business growth. Do you want to learn more about the power of data and predictive marketing for your marketing strategy? Check out this insightful article by Gary Stubbenhagen, our Head of Data at Builtvisible or redirect to our data product page to learn more.

Join the Inner Circle

Industry leading insights direct to your inbox every month.