How many of these comments and issue sound familiar?
Our executive leadership team gets conflicting data / reports!
We spend HOURS tracking down the "right" numbers!
What data sets are available?
Should I build my report in SFDC or use the data from the data warehouse?
How can I get access to the support data?
Who owns the Sales Pipeline dashboard?
There multiple definitions of MRR - What is the "right" calculation?
I need to report an issue with the Account data - how do I do this?
We are planning a Tableau update. Who are the stakeholders we need to contact?
I heard there is an outage on our Analytics server - what's the status?
I JUST heard about the change to a new Account Based Marketing dashboard happening NEXT WEEK!
I've developed a Customer Journey Dashboard - how do I let people know?
Our stakeholders stop using the dashboards we create for them after a couple of months!
Some of our stakeholders seem happy with the analytics team. Others, not so much. But we don't really know who is unhappy or why.
I've heard we are moving away from Tableau. Is this true?
What is our ML/data science strategy?
We struggle to effectively communicate all the analytics activities and success stories to our stakeholders and leadership.
We could go on and on, but you get the idea.... and hopefully, the majority of these comments/issues are not ones you hear very often.
If you hear these comments/issues more often than you like to admit, don't worry - they are universal. We have seen and heard them from 200 person startups to $10B companies.
From the outside, your analytics organization may seem to have everything it needs and is humming along: a robust data warehouse and data pipelines, powerful data viz tools, solid data governance, and a data science function generating useful models.
But the reality is you are likely still facing these (and many more) issues. These issues result not from your people, organizational structure, or tools/technologies.
These issues result from a lack of Analytics Operational maturity.
Sometimes they are merely minor annoyances. However, over time they will completely undermine the entirety of your analytics efforts.
Analytics Operations - The Missing Piece in the Analytics Puzzle
In a nutshell, Analytics Operations (not to be confused with Data Operations) - is a critical set of activities that include these elements:
Typically, Analytics Operations activities are handled by various people in the organization - often spread out across the Analytics team, IT, PMO, and Business Ops.
They are likely using a data governance tool to track metrics and definitions, a project management tool to track cross-functional analytics projects, a ticketing system for data access requests, and email to blast out maintenance and outage notifications to whomever they think needs to know, and spreadsheets to handle everything else.
None of it is centralized from an ownership perspective, a data perspective, or a workflow perspective.
Things fall through the cracks....
Trust the data and the analytics activities suffers....
Stakeholders are frustrated....
Data analysts are frustrated....
Executives are frustrated....
Eventually it will all catch up to someone. We'll leave it to your imagination to identify who that someone might be.
Improving your Analytics Operations
We advocate for the formalization of the Analytics Operations activities.
This doesn't necessarily mean a dedicated person or team. However, there does need to be a function - perhaps spread across multiple people or teams - to oversee these activities.
Nor, does the organizational alignment matter.
What matters is that this person, people, or team are utilizing a single set of tools, data, processes, and workflows to manage these activities.
To help assess the state of your analytics operations and identify areas that need attention...
Introducing the Analytics Operations Maturity Model
You are probably familiar with the traditional BI Maturity Curve - which shows the progressing analytics capabilities from descriptive to diagnostic to predictive and finally prescriptive.
Likewise, there exists other models (e.g. from Gartner) that captures analytics maturity - from basic to transformational - which focuses more on business outcomes and value instead of capabilities.
Not wanting Analytics Operations to feel left out in the world of maturity models, we have developed a framework to measure and track analytics operations maturity.
This model forms the basis of a more formal framework for organizations to assess their analytics operational maturity:
To fully assess where you land on the Analytics Operations Maturity Curve, and identify the specific areas that need attention, a survey of your stakeholders would be necessary.
Some sample questions for your analytics community (stakeholders, analysts, etc.):
Do you receive timely notification and updates regarding data/reporting outages and maintenance activities?
Do you know the processes for requesting access to key systems, data and dashboards?
Do you have visibility into the analytics project roadmap, status, and timeline?
Do you have a good understanding of all of the available data sets, analytics tools, and capabilities - and how to fully leverage them?
Are you sufficiently involved in the planning and execution of changes to the analytics environment that directly impact you (e.g. new tools, new data sets, updates to systems)?
Do you know where to find the official metrics/KPI definitions and calculations?
Do you have a good understanding of the data and analytics strategy?
Do you know who to reach out to if you have questions about the data strategy or analytics capabilities?
Do you know the process for raising data quality concerns?
A thorough survey would ask more - and perhaps different - questions than the above sample to get a complete picture of the existing analytics operations.
Done well, Analytics Operations is the "glue" that holds together all of the analytics activities.
Not done, or done poorly? The analytics activities lack cohesion and the organization is unable to fully leverage the power of their data.
Commentaires