Marketing Analytics vs Product Analytics

The established market definitions for Marketing Analytics and Product Analytics seem differentiating enough:

• Marketing Analytics platforms and tools collect, analyze, model, and visualize marketing data. They allow marketers to optimize marketing and advertising campaigns by better understanding prospects and customers, and their behaviors across channels.

• Product Analytics tools are used by product and growth teams to analyze and understand how users interact with their products or services. These tools help companies gather data on user behavior, such as what actions they take, how frequently they use the product, and what features they prefer. The data is then analyzed to provide insights that can be used to improve the product, make data-driven decisions, and optimize business outcomes. 

However, Marketing Analytics and Product Analytics appear to be on a collision course. A convergence seems imminent as:

1. Business users across traditionally disparate functions, such as Product, Growth, and Marketing, become more aligned around strategies such as product-led growth (PLG) and seek common reporting
2. Analysis of customer journeys relies on user event data, which can be captured by a single tool (e.g. Snowplow, RudderStack or Segment) for visibility across channels, and both frontend & backend layers
3. Modern data architectures store that event data directly in the data warehouse, along with data from other business systems

Historical Differences

The origins and evolution of Marketing Analytics and Product Analytics have resulted in the following assumptions: 

While these differences are legitimate, are they a result of historical evolution of these products, or is there any fundamental reason for them to be different?  To probe further, let’s ask the following two questions:

1. Is there an underlying architectural difference in the platforms for marketing analytics and product analytics?

The current answer is an emphatic ‘No’. Unfortunately for first-generation tools like Adobe Analytics or Amplitude, there are, because they were purpose-built solutions for marketing and product analytics respectively. They cannot be effectively used interchangeably. 

But what if you started from scratch as a generic event data analytics platform and did not have an opinionated data model, specific to things like campaigns, page visits, channels, ads, users, sessions etc.? 

This is what we did at NetSpring with our Relational Event Streams technology. It allows for representing and computing over both relational and event-oriented data models. The result is a platform that can foundationally serve both marketing and product analytics equally well. At NetSpring we believe that we are headed towards a convergence of marketing analytics and product analytics, but that is going to come from next-generation vendors like NetSpring that have built a generic platform.

2. Is there business value in a converged marketing and product analytics platform?

The answer to this is an overwhelming ‘Yes’. Marketing and product analytics tools have historically served different groups within companies, with no overlap. But that is not the case anymore with product-led growth models that are becoming the primary business methodologies for many companies. A marketer cares not just about acquiring customers, but also their engagement, retention and life-time value, to make sure the marketing spend is resulting in the most profitable customers long-term. A marketer may want to target campaigns at cohorts of users based on their behavioral patterns within the product. A product manager cares not just about usage of new features but also impact of those features on acquisition, revenue, and support. 

With product becoming the center of business, the lines between Marketing and Product (and also Sales and Support) become blurry when it comes to analytics. A converged analytics platform results in companies moving from siloed marketing metrics and product metrics, to more holistic business metrics. Analytics then becomes more business impactful because they are able to study the entire customer journey and experience from acquisition, through engagement, to upsell.

Pros and Cons of Generic Analytics Platforms

The advantage of a generic analytics platform is that it can be used to model a variety of use cases. You avoid multiple, disconnected tools – each for one aspect of analytics. It is scalable and can serve you for years as your business evolves. The sophistication of analytics you can do is high. 

The disadvantage of a generic analytics platform is that you have to model your use case using an abstract modeling layer. While the long-term benefits are great, it is not as easy to get started. Pre-packaged, purpose built solutions with an opinionated, rigid data model are easier to get started with. For small or less analytically sophisticated companies, a generic platform may be overkill.

Layered Approach to Analytics

At NetSpring, we provide a generic analytics platform. It has an abstract modeling layer and a language called NetScript that can be used to express analytic computations of arbitrary complexity. But we recognize the difficulty of getting started easily with a generic platform. To address this, we provide use case specific templates for the typical analyses that users often start with e.g. event segmentation, retention, engagement, funnel, paths, attribution, etc. Note that while these templates are specialized, the underlying data model is still generic. For instance, you can study the flow of a User just as easily as a Ticket or a Document, using a flow template.

The Modern Data Stack

Major shifts in analytics tools are often accompanied by major shifts in data architectures. Today, we are witnessing a shift to the Modern Data Stack. 

Cloud Data Warehouse

At the center of the modern data stack is a cloud data warehouse like Snowflake or BigQuery. These data warehouses are the central repository of all data in modern enterprises. This implies:

• Event data like product instrumentation events or IoT sensor readings, that traditionally never made it to the data warehouse, are now being stored there.
• Data from the thousands of marketing SaaS services companies use today are easily made available in the data warehouse with modern ELT tools. 
• Marketing products have themselves started providing tighter (even bi-directional) integration with data warehouses. 

Composable CDP

The second shift is the emergence of the composable CDP. Composable CDP means using best-of-breed systems centralized on the data warehouse. This implies:

• Specialized instrumentation systems like Rudderstack, Segment, or Snowplow that are decoupled from analytics tools, providing data in neutral formats in the warehouse for anyone to easily consume. No more critical customer data going off to some black hole SaaS service.
• Specialized ELT (as opposed to ETL) tools like Fivetran bring data from marketing systems like Marketo, Pardot, HubSpot, and Google Ads to the warehouse.
• Warehouse-native data transformation tools like DBT to transform raw data into more consumable forms. 
• Data activation to marketing systems such as Braze or Klaviyo directly from the warehouse.
• Warehouse-native marketing and product analytics sitting directly on top of the warehouse. No data leaves the warehouse ever. No additional ETL / reverse ETL mess. 
• Higher levels of security and governance by virtue of being warehouse-centric.

Conclusion

The distinction between marketing analytics and product analytics tools is historical, and will go away. While traditional vendors such as Adobe and Amplitude are likely to move in this direction, truly innovative products are going to come from next-generation vendors like NetSpring, that are optimally architected for this convergence. To learn more, request a demo today.