Applying Data Quality and Machine-Learning Basics to Improve Customer Experience and Reduce Cost
As a global media brand, our customer was struggling to engage their customers and prospects to clean their data and keep it up-to-date. This led to invoices not being paid, difficulty to contact and request consent for different marketing actions and other
To resolve this issue, an annual data quality budget was allocated so that offshore and onshore resources could manually solve the critical issues for the most profitable sectors. Of course, this addressed only a part of the problem and not the root cause of it.
Then, Data Trust Associates were introduced into the scene, and we were engaged in solving the cause of the problem with a limited budget and within a narrow time frame.
After a brief assessment of the situation, we quickly understood that the source of the problem was due to the combination of outdated legacy systems, incoherent business rules, and lack of smart customer engagement. Alongside these, we also observed that an organisational change was required to pave the way for a new working strategy.
With an aim of automating more than 85% of the manual work, we got the Board’s approval to implement the following plans:
- Define a Master Data source.
- Define and apply clear business rules.
- Use publicly available and legitimate open data for cross-verification.
- Apply machine-learning basics to the new data training model.
- Update legacy systems to incorporate new business data rules.
- Implement a data governance structure for the CRM data involved.
While several of the outlined steps are not rocket science, their implementation led to immediate success and buy-in from management.
140% Return on Investment within one year
The annual budget that was previously allocated for manual data quality activities was redirected to the DQ automation project. This was enough to implement the key changes so that 90% of the work is automated.
Rather than focusing on key segments and data entities, the implementation was effortlessly deployed to all data segments resulting in an immediate impact in the customer interaction experience.