Business Intelligence vs. Product Analytics: Why You Need Both

Data is the most valuable asset of our time. But organizations of all sizes are struggling to tap into the full potential of the information that they collect.

Part of the challenge is that the language of data isn’t always intuitive to our very human minds. So, we rely on software, spreadsheets, and statistical packages to make our best judgment calls.

In this article, you’ll learn about a new genre of customer analytics capabilities that speak the language of people, rather than systems. Let’s start with an overview of the modern business intelligence (BI) system — likely what you use today. 

The role of modern Business Intelligence

Business Intelligence has always been about creating faster, better, and more accurate processes for turning fragmented insights into prescriptive and actionable stories.  

With the launch of Amazon Web Services (AWS) EC2 in 2006, computing power reached an unprecedented level of scalability. As a result, SaaS BI platforms have become so efficient, accessible, and cost-effective that any company, of any size, has the potential to become a data-driven operation.

It’s this hunger for information that has given rise to the variety of BI solutions on the market today — each with their own value propositions.  Every company in the world needs data to strengthen its competitive footing. We live in accelerated times, and the companies that are winning the race for market share are the ones with the best understanding of their metrics. 

If you use Looker, for instance, you probably rely on this platform’s data capabilities. If you use Tableau, you’ve likely come to appreciate this platform’s extensive visualization library.  No matter which BI you use, your solution of choice is likely your right-hand in empowering the rest of your organization to make their best judgment calls. 

But you’re likely also hitting limits. Reports pique people’s curiosity, so teams start asking for more analyses. How can your organization keep up?

The limitations of SQL

SQL is the basis for many modern BI systems — including Looker and Tableau.  A lot of people know SQL, which is why so many BI companies default to it as a database technique. 

Keep in mind, however, that SQL is 46-years-old and became ubiquitous before the invention of the elastic cloud.  Back then, to answer questions, analysts had access to a tiny fraction of the processing power available today. Very few companies even had access to data, in the first place. It made sense to wait weeks, or even months, to run analyses. 

But today, businesses need to answer questions in seconds or minutes — not days or weeks. Consumers have sophisticated digital lives that often span multiple devices. A question that your company can’t answer is one that your competitor can. 

SQL can only take you so far because it speaks the language of databases.  Let’s say that you want to see how many people bought Brand X flavored water in New York City. That’s a straightforward descriptive analysis that SQL operators can handle.

But human life isn’t straightforward. Our purchase decisions involve a series of complex judgment calls, with many twists and turns. Let’s say that you want to know which customers bought Brand X flavored water and then purchased Brand Y flavored pretzels. Then perhaps a week later, they purchased Brand Z snacks.

These types of recursive queries don’t perform well. They often become challenging — and often impossible — to articulate using SQL. 

So how would you answer this question?

Most likely, you would export a dataset that you would re-import into a statistical package. From there, you’d run your analysis and export the results that you would re-import back into your BI.

That’s an arduous process and a roundabout way to get an answer to a simple question. Because it requires so many steps, it’s also prone to human error. It’s overhead that your very human mind doesn’t need.

Enter customer analytics

Customer Analytics enables users to answer complex questions in a more interactive way, without the use of SQL. In doing so, you save time and resources while uncovering substantially greater numbers of insights.

The elastic cloud has given way to the accessibility of scale. As mentioned in the paragraphs above, the limitations of SQL are what it can articulate, not the scale of the data that it’s using to perform analyses. That’s where customer analytics enters the picture — you can make analyses without needing to know the language of formal statistics.

Customer analytics platforms combine different types of events and actions to build a comprehensive picture of a user journey. That means you can see the exact paths that a user took from one place to the next, to accomplish a specific goal. You can analyze momentum-drivers through a conversion funnel in a way that you can’t with SQL. Customer analytics are in tune with the natural ways that human minds ask and answer questions in the moment.

With customer analytics platforms, companies can democratize access to data around a universal source of truth. With this common basis of understanding, people can work together in more collaborative and creative ways — fostering a culture of teamwork and innovation. 

Everyone can:

• Run the analyses they need, in a user-friendly interface, using the same universal source of truth that technical teams use
• Ask thoughtful business questions and seek answers on their own, without needing to learn SQL or relying on time-consuming report-generation processes
• Gain access to a user-friendly yet simple interface that empowers making judgement calls, consistently
• Democratize access to actionable insights, so every team can build their own quantitative processes
• Explore data in a way that is unique to our innate, human curiosity, without getting stuck in code

Example customer journeys

What does the concept of customer analytics look like in action? In this section, we’ll walk you through a few examples of the concepts that we discussed above, from the perspective of the Indicative platform. As you browse through these examples, try to imagine what the process would entail to create these with traditional BI tools — and likely, the addition of technically trained people using statistical software and programming languages. Very few team members outside of the data science or business analyst function would be able to run these analyses.

To provide context, these analyses are based on a dataset for a hypothetical company: a subscription-box company that sends you new products every week for your kitten (we hope this actually exists, somewhere), KittyCart. We’ve included some screen recordings that you can browse and easily pause for a closer view of the concept.

Let’s say that you want to visualize typical customer journeys.

In this situation, you would use the funnel analysis tool within Indicative. The tool makes it possible for you to segment audiences, at the moment, to see what various user flows look like within your product. 

Here, you can visualize the different ways that users are navigating from a site visit through subscribing and then further engaging with your product. You can see, for instance, the difference in conversion rates by various types of marketing source, over a time span of your choosing.  You can even see details pertaining to when someone entered a particular funnel and how long it took for a conversion event to take place. 

Now, you’re likely curious about the business drivers behind the trends that you see.

There’s always a “why” behind the numbers that you see in your analytics. Numbers, alone, may not uncover the full picture of why you notice a particular trend. With annotations, you can bring more context to the quantitative story that you observe. 

Let’s say that you launch a marketing campaign, partnering up with local pet sitters or running a series of billboard ads in a local market. Or maybe, you’re thinking about launching new subscription tiers. 

Just add an annotation, so everyone at your company is on the same page about what’s happening and why. Then, anyone on your team has access to this contextual information when analyzing the trends that surface from your dataset. 

How can you take action with the insights that you’re generating?

Your data is only as valuable as the feedback loop you’re building back into your business. So what’s the best way to make use of all the information that you’re gathering?

The short answer: narrow down the segments that are most important to your business. One way that you can build these segments is to group audiences by a specific set of actions that they took to engage with your business. You can then monitor changes to this segment over time.

Let’s say that you want to gain a better understanding of steps that a website visitor took to become a subscriber. In the video above, you can see an analysis that groups visitors by whether they viewed a particular website feature. In this example, the feature is “a kitty cam,” a part of a more general website conversion funnel, that shows a live-feed of kittens having fun (sidebar: here’s a pretty awesome one from National Geographic).

This perspective will help you understand whether specific features within your user experience are performing as intended — and whether certain features are likely to spark higher conversion rates than others.

Final Thoughts

Everyone at your company needs access to data to make their best judgment calls. But data, alone, isn’t enough to give your company its best strategic advantage for understanding the needs of your customers. Your technology stack needs to do more heavy lifting to help you transform quantitative insights into human stories. SQL can only take your team so far. To take bigger jumps forward, reconsider and reimagine what’s possible with your database.

Customer Analytics is an essential part of the technology stack that enables individuals to make real-time ad-hoc analyses without coding or SQL and outside of the direct support of your company’s data engineering team.

Contributors Statement
This work was a collaboration between the Indicative team. Jeremy Levy, Esmeralda Martinez, Marc Liebmann, Tara McQuaide and Bjorn Sigurdsson contributed to the narrative.

Disclosure
The opinions expressed are those of the authors. This material has been prepared for educational purposes only.