It is clear that data teams are now essential in most organizations, and it is paramount to measure the value that they bring. Although it can be difficult, it must be done. So, here are some tips on which KPI to choose for the data team on three levels: engineering (data engineers, data architects and DataOps), consumption of data (data analysts, data scientists), and governance (the CDO).
Regarding data engineers, data architects and DataOps, KPIs must be focused on their ability to build an effective data platform. That means establishing operational KPIs around the following themes:
- The data platform : these are KPIs that allow for the measurement of elements forming the data platform and its mode of operation, thus validating the work done by data architects and DataOps. We can use KPIs like the number of errors and incidents, the error resolution speed, the automated production deployment success rate, the reuse rate of tools and frameworks,…
- The data quality: these KPIs allow for the validation of the data in terms of collection and provisioning requirements. Some criteria, like accuracy, integrity, consistency, security and up-to-dateness, areiinherent to the data, while others, like relevance, reliability, usability and conformity, are extrinsic and depend on the business project.
- The data lifecycle: like the name implies, these KPIs allow for the validation and measurement of the data lifecycle management process. They include data uptime, board response time, data refreshment frequency Le cycle de vie de la donnée : comme son nom l’indique, ces KPI permettent de valider et de mesurer le bon fonctionnement du processus de gestion du cycle de vie de la donnée. Ils comprennent ainsi le data uptime, le temps de réponse du board, la fréquence de rafraîchissement de la donnée, le taux d’usage des data products par les end users,…
With these KPIs, the data team will be able to measure itself and enter a virtuous cycle of continous improvement of quality, performance and usability of the data platform.
The data consumers
As data engineers and architects build the foundations of the data platform, data consumers make it a tool that directly brings value to the business teams, thus being the other side of the data team’s value proposition.
Data consumers are the second big family in the data team, and are divided in two groups: data analysts and data scientists.
Those two groups have KPIs in common, mostly regarding project management. Those KPIs include, inter alia, the number of projects done, and the fulfillment of delays and engagements. However, each group also has KPIs that are specific to their expertise.
Data analysts are measured on their time to insight, which is time the time they take to generate an analysis when they are asked a question, and the time to action, which is the delay between the generation of their analysis and the decision taken by business teams.
On the other hand, the efficiency of data scientists is measured by the effectiveness of their models, which entails their accuracy and their deployment delay.
The data team governance
Like any business unit, KPIs exist on the data department scale, to measure the budget management aspect on one hand, and the expertise and intrisic value of the department on the other.
Financial KPIs help evaluate the budgetary performance by correlating the budget and the realizations done at the end of the year.
From an operational standpoint, department governance KPIs must measure how data exploitation allowed for the fulfillment of the company’s strategic goals.
We can define a matrix intersecting the different strategic stakes of the company and the different business units (see picture below). This way, the CDO will be able to map out BU by BU or stake by stake data’s contribution. He can use different methods to empahsize it, like a point system or a calculation of the ROI attributable to the data team.
In general, KPIs aim to measure the operational aspects on one hand, and the budget control aspects on the other. Regarding the data platform, operational aspects cover the relevance and the efficiency of the platform, and the other aspects verify that data exploitation brings value to the different company departments.
In conclusion, measuring the data team’s ROI is crucial to optimize the strategy and begin a virtuous cycle.
At Indexima, our expertise is the data platform. That’s why we have created Autopilot, a solution that fires up your Snowflake performance. Wanna know more ? Click here !