TLDR: Disparate data occurs when teams pull, categorize, filter, and interpret data in unique ways, resulting in different numbers and conclusions.
Few things are more paralyzing for your business than data that doesn’t add up.
Bad data taxes your time, reportability, revenue, and cohesion. If your workplace lacks a holistic approach to systematizing data that teams universally follow, this Tough Talks Made Easy is for you. You’ll learn to explain to your CMO and CTO why disparate data occurs and how to gain clarity through data unification.
The roots of dirty data
Organizations at one of two stages of maturity tend to find themselves with dirty data:
- startups that have scaled fast after receiving funding, and
- legacy enterprises with a mix of offline and online data storage.
Startups that prioritized growth with limited resources have likely been too busy to explore strategies for systematizing data. While teams in more mature organizations have had time to develop siloed, well-entrenched methods of treating data.
In either case, there’s a lack of thinking about data holistically or creating agreed-upon systems and practices. And when it comes time to make key financial decisions (e.g., how to utilize marketing investments next quarter or year), discrepancies ripple through the processes of surfacing data.
No centralized repository
When teams use data transformation tools to visualize their numbers, data yields can wildly vary without a centralized repository for calculations.
For example, let’s say marketing and sales use different standard and customized filters to denote types of opportunities, leads, and revenue.
Teams might use different datasets from their CRM vs. their marketing automation platform to answer the same business question and further obscure the analysis by classifying comparable fields in distinct ways (e.g. region vs. country).
No shared understanding
Around the business, departments might use particular tools and systems that allow for reporting to varying degrees of sophistication.
Let’s assume your business has an agreed definition of Conversion: the movement of a person between the Inquiry and Active Opportunity stages of your database.
Your leaders must ensure that different teams have a shared understanding of calculating this. Trouble occurs when:
- one team uses Tableau to calculate conversion rates using a cohort analysis — looking at a selection of your audience that shares a behavior, and
- another team uses Salesforce to divide the number of MQLs by the number of opportunities per calendar month.
The time it takes for a lead to go from an MQL to an opportunity is calculated per number of days. Unless the velocity of leads is extraordinarily fast, converting to opportunities in a number of days, you won’t calculate an accurate conversion rate by dividing the number of MQLs by the number of opportunities in the same month.
Let’s say you have 50 opportunities and 100 MQLs active in October.
Taking a monthly view would suggest that your conversion rate is 50%. But there’s no evidence that these opportunities came from October’s MQLs. By contrast, a cohort view would allow you to look back and see of those people who had MQL status in July, who became an opportunity since then?
Alternatively, if you have 50 opportunities in October as a cohort, you can see the 30 MQLs who resulted down-funnel in opps, and these are the parameters you can use to accurately calculate conversion.
The bottom line
You get disparate data when teams pull, categorize, filter, and interpret data in unique ways, with particular mechanisms for reporting and analytics.
Let’s bring it together.
Enter data unification
Data unification requires a maturity of thinking about and organizing your data. Culturally, it means acknowledging disparate data as a systemic problem in your workplace.
If leadership isn’t convinced that addressing this is a priority, punctuate the range of competencies that disparate data puts at stake: reportability, decisiveness, time, cohesion, revenue, and clarity. People across the organization gain those benefits from data unification—a compelling basis to incentivize people internally.
The best practice for data unification
Think of your data as holistic, within a clearly defined and universally adopted structure. This is both theoretical and practical.
For example, different departments need to agree upon and reference the same definitions in their analytics and reporting—they shouldn’t use their own methods to filter and calculate things.
Streamline your data storage by pulling it from decentralized repositories and siloes into a data warehouse. From there, build out a central repository of definitions and calculations that everyone then references (e.g. all teams adopt the same definition of sales velocity and reference the same cohort structure and conversion math).
With regards to tech, it’s sensible to consolidate the platforms your organization uses to analyze, store, transform, and visualize data. This ensures teams can access, surface, and parse data consistently and narrows the scope for discrepancies. Learn more steps by reading our post ‘How To Create a Data Hygiene Plan That Works.’
Looking forward
Disparate data is not an insurmountable problem, but it’s more insidious to fix the longer you leave things broken.
If your workplace has lacked the resources or foresight to structure data holistically, or you’ve sat in one too many meetings where different teams make the case for their numbers, now’s the time to act.
Data unification is a project that will strengthen the integrity of your data and decisions and help to drive your business forward.
Get in touch for guidance on systems and practices that make data work.
