When I announced that I had joined Hightouch, several folks messaged me asking how Composable CDP products like Hightouch compare and/or complement digital analytics. That makes sense since I have spent the last twenty years in the digital analytics field! In this post, I will elaborate on one of the reasons I joined Hightouch, which I mentioned in my introductory blog post:

Digital Analytics Isn’t Enough - While digital analytics tools are great at improving websites and mobile apps, this data must be combined with all other customer data to enable true personalization and experience optimization. Organizations have complex data that goes far beyond behavioral data. Because of this, I have seen many digital analytics teams begin sending digital analytics data directly or indirectly to cloud data warehouses, where it can be part of a 360-degree view of the customer.

Why digital analytics alone can go wrong

I’d like to start by being very clear - I am and always will be a fan and advocate of digital analytics! I have devoted a significant portion of my life to helping organizations get the most out of their investments in digital analytics. Collecting and analyzing behavioral data from websites and mobile apps is essential to improving digital products and experiences.

However, there are situations in which only relying on digital analytics could lead to worse customer experiences. You read that correctly. Sometimes, digital analytics can cause more harm than good. Why? Digital analytics tools often contain only a subset of customer data, and your digital analytics team may [unintentionally] hinder your marketing team’s efforts to deliver excellent customer experiences. But, combining the rich behavioral data in your digital analytics platform with other customer data can up-level your digital experiences. Recent advancements in cloud data warehouses and Composable CDPs (like Hightouch!) are helping augment what can be done in digital analytics tools. To illustrate the problem of relying solely on digital analytics, let’s look at a hypothetical example.

Imagine you work for an eCommerce company and run a digital campaign to attract customers. One of the customers you attract from an advertisement visits the website and places a product in the online shopping cart but cannot complete the purchase due to a technical website problem. The customer then engages in an online chat with your customer support team (e.g., ZenDesk). Still, they cannot fix the issue immediately, so the customer is told the company will follow up via email when it is resolved.

Next, imagine that a few days go by. Your marketing team uses your digital analytics tool (e.g., Amplitude, GA4, etc.) to build an audience of users who have abandoned products in the shopping cart and sends this user audience to its email provider (e.g., Salesforce Marketing Cloud, Braze, Iterable, etc.) for activation. Cart abandoners receive a reminder that they have items left in the cart and may even receive an offer of a slight discount as an incentive to complete the purchase. But if this particular customer still cannot purchase the product, the email only reminds them of the frustration that the website doesn’t work, and the price discount makes them wonder if they have been overpaying in the past!

Next, the marketing team, a big believer in “customer journey optimization,” has a pre-built customer journey setup that identifies customers who still have not purchased products abandoned in the cart after fourteen days. Using the digital analytics tool, they have a second pre-built cart abandoner audience sent to Google and Facebook for digital advertising. Now, the customer who is unable to purchase the product they wanted is being shown (taunted by?) the product on virtually every website they visit!

Thus, instead of using marketing technology to build better customer experiences, it’s likely that it has turned off this customer and possibly sent them to a competitor! And worse yet, your organization is spending money on digital advertising to target a customer who already wants to buy the product but cannot. If, by some miracle, the customer eventually purchases the product, you likely have lost money due to an unnecessary discount offer and unneeded advertising spending.

So why did this customer experience go so wrong? The root of the problem is that the company was only looking at a subset of customer data in digital analytics when it built and activated customer audiences.

Amplitude cart abandonment cohort

From a digital analytics standpoint, nothing was done wrong. A user abandoned a product in the shopping cart, so they were emailed a reminder and a discount. When that didn’t work, they were pushed ads to entice them back to the website to complete the purchase. That’s how digital analytics and digital marketing are supposed to work. However, because the digital analytics tool wasn’t aware of the customer support issue impeding the customer’s purchase, it inadvertently delivered a sub-optimal journey. The lesson here is that you cannot provide effective customer journeys if you don’t take into account all customer data. Sometimes, you have to up-level digital analytics and view it as part of a larger customer data ecosystem.

How cloud data warehouses and composable CDPs can help

How could cloud data warehouses and a composable CDP help the digital analytics team in this situation? To avoid a negative customer experience, we need website/app behavioral data, CRM data, customer support data, campaign data, and email data for all customers in one place. Most organizations will deploy a Customer Data Platform (CDP) to accomplish this. CDPs were built to ingest data from multiple tools, unify user identities, and build audiences that span all customer data. In this case, the company should have built an audience of customers who abandoned products in the cart but excluded any customers with open support tickets.

Hightouch cart adandonment without support ticket audience

The simple addition of customer support data would have prevented the customer from turning from an advocate to a detractor. However, building the correct audience is impossible with digital analytics tools since they only track part of the customer experience. In contrast, creating this type of audience in a CDP is straightforward.

While this explains why a CDP is needed, why would this company want to use a Composable CDP instead of a packaged CDP? Packaged CDPs require you to send your data - digital analytics, customer support, email, etc. - to their proprietary data warehouse. Combining all of this data in a packaged CDP can be extremely time-consuming (6-12 months), and since most CDPs charge by data ingested, it can be extremely costly (💰💶💸). Consider how much data you have in all of the SaaS tools your organization uses. The price tag of sending all SaaS data to a packaged CDP would be enormous!

Conversely, since most organizations desire one place to contain all customer data, the industry has responded by providing cloud data warehouses like Snowflake, Databricks, BigQuery, Redshift, etc. Cloud data warehouses are highly flexible and cost-effective. Today, cloud data warehouses are the CDP.

Packaged vs. Composable CDP architecture

If your organization has aggregated its customer data in a cloud data warehouse, Composable CDPs like Hightouch provide traditional CDP functionality without requiring you to duplicate or create another silo for your customer data. Identity resolution and audience building can take place natively from your cloud warehouse, and audiences can then be sent to SaaS products for activation. The Composable CDP approach allows organizations to capitalize on the investments they have already made in their cloud infrastructure.

In the preceding example, a Composable CDP like Hightouch could be used to build the appropriate audiences using data from digital analytics (Amplitude, GA4), customer support (Zendesk), and email (SFMC, Braze, Iterable) to ensure that each customer is placed into the correct marketing journey/activation flows.

This customer scenario is just one example of why data activation is much more effective when it leverages all customer data and why many organizations have started leveraging the warehouse-first model, which runs activations directly from the warehouse. You could imagine many other scenarios:

• A customer leaves an item in the shopping cart but ends up purchasing the product in a local store or by phone, but the digital analytics platform isn’t aware of the offline purchase.
• A customer with a high lifetime customer value but who hasn’t purchased much online may not be treated online as a “VIP” because the digital analytics audience only sees part of the customer’s history.
• Online personalization strategies can be underutilized or mistaken if they are based only on data contained within online behavior. For example, a customer may buy shoes online, but different products in-store and online personalization may suggest the wrong offers or incorrect next product to purchase.

Summary

As stated at the beginning of this post, digital analytics is essential to improving digital experiences. The combination of marketing, product, and experience analytics helps organizations identify where customers are struggling and opportunities for improvement. Identifying and fixing digital experience issues can lead to incredible ROI and better customer satisfaction.

But there are times when digital analytics data needs to be up-leveled and merged with other data sources to provide better multi-channel customer journeys. This desire for comprehensive customer journeys is why organizations are turning to cloud data warehouses as the ultimate repository of customer data and to Composable CDPs to help turn that customer data into increased customer value and cost savings.

If you’d like to learn more about this topic or chat with me, please book a meeting with me.