Are you REALLY a data driven organization?

When the discussion turns to data and analytics, the phrase "data driven organization" is sure to come up.

It may be in the context of the speaker claiming that their organization is data driven, but more often than not, that they aspire to be.

Either way, it is on everyone's mind.

Most surveys on this topic find that the majority of organizations say they are not fully data driven - including the 2020 Executive Survey by NewVantage where only 38% of the respondents report that they have a data driven culture.

But are we all on the same page as to what it means or what it looks like in the real world?

Well, quite simply... No.

Often, the discussion on this topic revolves around the ability to create dashboards and predictive models - and whether or not the speakers organization has the perceived ability to quickly generate and share them with the business. It is as if the act of creating and publishing a dashboard is evidence of being data driven.

Other times the conversation centers around the fact that the speakers organization is not data driven because they lack the necessary tools to generate and publish powerful dashboards and models.

In our experience, there is actually a very specific and consistent set of characteristics that are common across organizations that are making the most out of their data and generating meaningful insights that are driving business decisions.

When engaged in conversations on this topic, I will pose five seemingly simple questions that can take 60 seconds to answer, but often much longer to explain. Are you a data driven organization? Let's find out....

• Leadership trusts the data and metrics that are produced (Y/N)

• Insights able to be generated at “The Speed of the Business” (Y/N)

• There exists a robust framework through which insights are cross-functionally generated and consumed (Y/N)

• ALL of the data available within the organization is accessible for analysis and being leveraged to generate insights (Y/N)

• The consumers of the analytics have a well-defined and consistently applied methodology for incorporating information into their business processes and decision making (Y/N)

Let's take a closer look at each of these questions...

Trusted data?

Ultimately, analytics is founded on trust. Trust that the underlying data is complete and accurate. Trust that the systems, tools, and processes that collect, aggregate, and store the data are working properly. Trust that the metrics and insights are being generated in a consistent and reliable way. Trust that the insights are being interpreted and actioned correctly.

That is a lot of trust going on! To be honest, I'm not sure there are many organizations where there is complete trust in all these areas.

However, there is a simple - in concept at least - way to (re-)gain trust in the data and insights... by reporting on each step of the "analytics supply chain". This can be accomplished by developing and publishing:

• data quality reports on key data sets

• ETL jobs (success/failures/run time) reports

• a business glossary of metrics definitions and calculations that are leveraged by analysts

• business process maps that show where and how insights are consumed

Simple in concept, non-trivial to implement. However, if you want to be truly data driven, you cannot be wishing or hoping that the consumers trust your analytics. And the trust needs to start with leadership. If they trust the insights, so will their teams.... and with the above analytics supply chain visibility, any doubts that arise can quickly be addressed or squelched.

Insights generated at the speed of the business?

This can be achieved with a well-implemented ecosystem (the people, processes, and technology) that enables self-service analytics.

Contrary to what some stakeholders may say, not all analytics needs to be done in (near-)real time. On the other hand, having to wait 4 weeks to add a new SFDC object into the data warehouse so the dashboards can be updated is not "at the speed of the business".

Here are a few recommendations for developing the capability to generate insights at the speed of the business:

• Implementing a near-real time database where a small subset of critical data is frequently refreshed.

• Providing the capability for "citizen data engineering" - where business stakeholders can use data aggregation tools (e.g. Alteryx) to "join" their own data with official data sets from the data warehouse. Managed properly, this can result in the data engineering team taking the work done by the business and productionalizing it - with a significant reduction of the back-and-forth discussions about what data is needed and how it will be used.

• Establishing an agile data engineering and dashboard development process where the analysts are involved in the QA process to speed up the UAT process

Cross-functional insights?

The natural course of analytics evolution within an organization is that each function establishes their own analytics team - using their own database. This happens due to the fact that the business need for data starts on day 1. A formal analytics team doesn't appear on the scene until the business teams are well down the road of doing it themselves.

Eventually, a corporate data warehouse is created, but invariably what gets created is isn't so much a data warehouse, but something closer to a repository of separate datamarts - where the old silo databases are migrated to the warehouse.

Hopefully, at some point down the road, a proper unified corporate data model is implemented in the data warehouse / data lake / hybrid. However, while the data model is now enterprise focused, the actual analytics often remains silod - where sales, for example, is doing their analytics independently from marketing, finance, etc.

And that is where things remain....

However, truly data driven organizations will have recognized the power of cross-functional insights. Sales is highly dependent on marketing data. Finance is highly dependent on sales data. Customer success and product are highly dependent on support data, and so on....

These organizations build a framework that not only enables, but encourages cross-functional insights.

Some ways that this can be accomplished:

• Building a set of common/generic "business views" that are used by all analysts in all analytics teams. So, when looking at customers for example, the sales, marketing, and support analysts are leveraging the same "table" with the basic business logic already built in.

• Centralized tracking of analytics projects. Providing the sales analytics team a readout on the marketing analytics projects helps to drive alignment of analytics, reports, and dashboards. Say for example that marketing responsible for sales stage 0, but then the lead gets handed off to sales after it becomes an MQL... So, what happens when marketing is changing the definition of MQL? Well, unless there is cross-functional visibility to this effort, sales will be in for a big surprise when marketing updates their dashboards with the new MQL definition!

• Establishing a role or position for stakeholder engagement. One responsibility for this role is to help "connect the dots" across the various analytics activities across the organization

Leveraging ALL of the corporate data available?

For various reasons some corporate data may end up not being available to the data analysts. This can happen due to technical, budgetary, legal, or even political reasons.

A key concept here is that of "Data Democratization" - that the corporate data belongs to the entire company, not held hostage by a team or department. Of course, there is some data that cannot be shared with the entire organization due to legal or regulatory constraints.

Let's look at some examples where access to data can cause "blind spots" within the organization.

• The Sales organization has built out their own datamart to support sales reporting. Sales leadership considers this datamart as the source of truth and wants to continue to own and manage the datamart even after the creation of the corporate data warehouse. Finance and other functions are able to access and consume the data, but only as business views - eliminating the ability to cross-check the data (which does not match the corporate data warehouse as it lacks the custom business logic in the views).

• Social media data (e.g. comments) from customers is housed in a separate database that can easily handle unstructured text data. Technical challenges exist in connecting this data to the corporate data warehouse creating a serious blind spot for the customer success team.

Another area where truly data driven organizations excel is in having awareness of all the available data - and a passion for leveraging it. This is where a data catalog comes into play - providing visibility into available data sets.

Framework for business consumption of data and insights?

One of the most painful things that happens in virtually all organizations is when reports, dashboards, and predictive model output are "thrown over the fence". The dashboard is requested by the business, the data analyst builds it, and the business uses it. For a few weeks. Then usage drops off never to return.

The reasons for this are many (to be explored in another article), but fundamentally the root cause is a lack of understanding (by the requestor and the developer) of how to integrate the information generated into a formal business process.

Data driven organizations understand not just what data and insights are wanted by the business, but what is NEEDED by the business. In non-data driven organizations, what is wanted (requested) by the business may not the same as what is needed - but what gets build is what the business requested.

So - how does an organization enhance (or develop, as the case may be) their ability to consistently and successfully integrate information into their business processes and decision making?

That would take an entire book (and there are some excellent ones out there) to provide an in-depth answer, but there are a few concrete actions to help your organization being moving in the right direction:

• Transition the analytics function from a ticket / data request fulfillment center to an analytics consulting function. This requires both the analytics team and the stakeholders to shift their mindset. This will not happen overnight, but make the commitment to begin the process - perhaps by having someone from the analytics team participate in business planning sessions for the upcoming fiscal year.

• Track and follow-up on previous requests for reports and dashboards. Most dashboards die off within a couple of months. Dive into the reasons. This will unlock a wealth of insights into how data is being consumed, what is working, and what is not working.

• Develop data/process flows for key business activities. If business processes are not documented (it's OK - most aren't), start with one or two key business processes that are highly dependent on data. Map the business processes and overlay the data flow. If there are existing documented business processes, revisit them and overlay the accompanying data flows. Share them with the data engineering team, the analysts, the stakeholders. Move on to the next business process.

You are now armed with some ideas to begin making inroads towards becoming one of the 38% of organizations that are data-driven.

Keep the keys to success close at hand:

The 5 Key Characteristics of a Truly Data Driven Organization

Good luck, and please share your challenges and success stories!