Re-balancing Priorities: Why Data is a Service Function and should not be only viewed as a Product.

I’ve been working in the Data and Analytics field for more years than I’d like to admit. I’ve seen a huge amount of change. Business Intelligence has come a long way in a very short time. There may be a handful of companies that still have one or two “data people” sitting in the back corner running some reports, but generally the industry has grown and will continue to grow.

Several years ago we saw a shift in the way output from data and BI teams were being classified. Treating Data as a Product makes a lot of sense. I agree with the concept. The thought is for data to generate more value for a business it must be repeatable and scalable. Once that is achieved then these processes can be automated and the data analysts or engineers who created this can move onto the next business problem.

If this process is not “productionised” then there is a finite cap in output which can be achieved by the resource. Once you reach that cap, the resource is not able to develop anything new and must maintain and manage the existing set of output, creating technical debt — be that data tables, machine learning models, business reporting and visualisations.

In order to generate more you must hire new resources. Which costs money. Highly skilled data engineers or data scientists are not cheap due to the specialised nature of the role.

By treating these assets as products you can scale and automate and then dedicate a smaller amount of resources to monitor and then fix issues only when they break.

I’ve seen this model adopted and work very successfully. Bringing data closer to Tech methodology has aided the growth of data. This success has rapidly demonstrated the incremental value that data can drive to a business and warranted the investments made in Data both talent and infrastructure to build on these platforms.

Where does this end though?

Data being treated solely as a product, where use cases and personas are condensed down and then delivered runs a risk of removing the human element of what data is capable of.

You only build a product when you know you have a market for it.

If you begin to treat data only with a product lens, you forget that data originated from the fact that there were some IT people who were storing information in a database. This was then able to be interrogated with business questions that people were asking.

For me data is a service function. It is there to support a business through a variety of means. It could be providing visibility of business performance via KPI’s. It could be investigation and understanding of business problems, answered through deeper analytics conducted on what data points are available. Or it could be serving as the incremental revenue driver through more advanced algorithms like recommendation engines powering personalised display carousels on e-commerce sites.

There is a risk of forcing all users down the product route and giving a generic menu of what they can use.

By removing that engagement with the end consumer of the data you lose sight of what business question they are trying to answer.

The value in what those partnerships create and how working collaboratively the problems are being solved with data through recommendations and informed decision making.

Not to mention all the caveats about the data availability and quality used in order to generate an outcome. Unless that is disclosed to the end users of the data product then interpretations could be misleading.

Removing ourselves from interacting with stakeholders’ and having product managers asking them to submit tickets and creating user stories undervalues the role that data professionals can play in an organisation. Reducing these touchpoints between the business and data experts also can create a problem through giving a false perception of where data comes from and what needs to happen for it to become usable.

The rise of Data Fluency is looking to help address this. There is risk when you disengage or distance yourself from the stakeholder or end users of your “products”. There is no opportunity for them to become more informed about the steps required to take data in a raw form and convert it into something which can answer business questions and generate value.

So — what’s my point I’m trying to make?

Treating data as a product is an effective method to achieve scale and value. Treating data as a service allows for understanding, influence and informing at a level which will also create value. By combining both aspects we can continue to drive the growth of data within a business by being able to scale as well as staying connected to the people we are supporting.

Let’s not be so quick to want to disengage from stakeholders. For me that’s where I get the most satisfaction from. Helping others see and understand data, to make a business more successful.