By Upuli de Abrew, Co-Founder and Director at Insight Consulting
As it is in the natural world, data also experiences a lifecycle, from gestation through birth, metamorphosis, maturity and eventual death. Just as a young bird needs to be taught to fly and a lion cub must be taught to hunt, data also needs care and direction from custodians. However, unlike nature, data will rarely be able to stand on its own and instead need constant tending from its custodians, known as data stewards, throughout the entire data management lifecycle (DML).
The data management lifecycle comprises data collection, data ingestion, data storage, data usage, data maintenance, data archiving and data destruction.
Effective data stewards spend a great deal of time ensuring that data quality is maintained throughout the DML. Data quality refers to data that is accurate, valid, complete, consistent, accessible, unique, timely, relevant and reliable. Ensuring effective data quality requires a comprehensive approach, integrating people, process and technology.
Turning the lens inwards: A mini-case study
In our consulting business, the efficient scheduling of consultants, and the tracking and billing of time spent on projects is fundamental to our business operations. Over the years, we evolved from manual time capturing in Excel, consolidated by administrators for billing purposes, to a custom-built integrated scheduling, time capture, approval and invoicing system, which feeds our data analytics.
However, despite the implementation of these advanced systems, we found that we were still failing to achieve the targets we had set for ourselves, and had issues with high volumes of non-billable hours, missed invoicing deadlines, large amounts of time spent by account managers discussing timesheets with customers, as well as fluctuations in over- or under-capacity of our consultants – all impacting on our bottom line.
On close investigation, we uncovered several fundamental problems related to our data quality:
-
Consultants were capturing their time weekly or monthly, meaning many could not remember how they had spent portions of their time and so they booked these as non-billable hours.
-
We were not enforcing a final capture deadline to allow for efficient month-end processing of timesheet extracts and invoicing. If consultants captured their hours for a month after the invoicing run for the month, these hours would not be billed.
-
Tasks on projects were not broken down to a detailed level, which made it difficult to analyse where time was being spent.
-
Projects were not created timeously, meaning that when consultants captured their time, rather than ask for projects to be created, they would capture their time against a similar project for the same client.
-
Consultants and team leads did not have ongoing visibility or tracking of their timesheet hours as timesheet analytics access was restricted to management.
-
There was no formalised scheduling mechanism – account managers would approach team members or team leads directly, assign consultants to work, and then realise only on project kick-off that consultants were over-committed.
-
The flipside of this was that some consultants would have work cancelled at the last minute by customers who weren’t ready, resulting in them being left with excess capacity that had not been planned for.
Realising that many of our issues stemmed from data quality processes over the data management lifecycle, we implemented incremental changes, focusing on all aspects of people, process and technology.
The first thing we did was to communicate with our consultants, to explain how the accuracy of timesheet capturing impacted our business profitability. We put in place a resourcing manager who worked with team leads, who in turn worked with consultants to ensure that everyone understood how their hours contributed to our business profitability and growth.
We provided access to timesheet analytics for all team leads and project managers for increased visibility, while team leads were asked to report back on a weekly basis about how their teams were tracking against billable targets.
Account managers were asked to create projects in the system as soon as they were approved by customers, meaning these were available to consultants to capture against immediately.
We created standard tasks that followed our development methodology for each key consulting area, and consultants were trained on how to capture their time accurately. Timesheet capture is now also a part of induction.
Once consultants had an understanding of how important accurate time capture was, they were encouraged to capture their time on a daily basis. We implemented a process that sends reminders to consultants if they have not captured their time in more than two days.
We also implemented the ability for consultants to capture their planned time, and team leads ensure that planning is captured for all team members as far into the future as possible. This ensures that we are able to quickly identify consultants who will have capacity in the near future and can therefore be allocated to projects. Planned time is reviewed on a weekly basis.
Customer projects are classified into “Pipeline”, “To be Allocated”, and “Allocated not Started” for scheduling purposes. Projects are only allocated to consultants and captured on receipt of official confirmation from customers. This prevents cancellations from impacting capacity.
Account managers are encouraged to perform regular data audits to ensure that billing for their customers is captured accurately, and that non-billable time allocated to customers can be explained.
A culture of data quality
Creating a culture of data quality starts with recognising that data is a valuable asset and that everyone in the organisation plays a role in maintaining its quality. In our business, our data quality initiatives are ongoing, and require us to consistently make incremental adjustments and improvements to our processes. Our focus on small actions resulted in delivering an outcome far greater than the sum of its parts.
We learnt a great deal about how to foster a culture of data quality. It requires leadership commitment, communication and awareness and ensuring everyone is a data quality champion. It requires visibility of data and ensuring there is an overall data custodian. In addition to this, there needs to be continuous training programmes, as well as instilling a culture of accountability and recognition, and ongoing feedback and improvement.
Data quality challenges can occur throughout the DML, and addressing these issues requires an approach that integrates people, process and technology. While implementing robust processes and advanced technology can make a difference, data quality cannot be achieved without a strong focus on people.
