Mahmoud Yassin, Senior Data Manager at Booking.com, speaks with Derek Strauss, Chairman at Gavroshe, and CDO Magazine Editorial Board Member, in a video interview about target architectures, building architecture from a platform perspective, analytics tools, and the importance of having KPIs like data quality score.
Speaking about target architectures, Yassin states that no fixed architecture fits all organizations. He adds that the target architecture depends on various factors.
The factors that can affect the target architecture include:
Organizational maturity
Data literacy in the organization
Standpoint of data and analytics
Business intelligence needs
AI needs
Platform strategy
Adding on, Yassin recalls that the architecture at his previous workplace was something between a data mesh and a data lake which worked well. Explaining further, he says it had governance by design.
For instance, Yassin says, it is like developing a data product and keeping the controls as a data owner or producer. Further, certain controls may also be induced on the consumption side by introducing data contracts wherein one would need to explain the need for data.
Yassin believes that asking questions like why the data is needed, the severity of data required, or the time frame of requirement also helps shape the architecture.
From the platform standpoint, there are many considerations to go through based on the needs, he says. Certain limitations come based on data processing and transformations due to the choice of platforms – on-prem or cloud.
There are numerous services on the cloud to choose from while working on-prem has its bindings, says Yassin. He stresses how one cannot leverage machine learning services from Azure or AWS while on-prem and must build their models.
Additionally, he emphasizes that metadata is key to the successful implementation of the data architecture.
When asked about the importance of analytic tools, Yassin denies the need for fancy tools stating that analytics can be done with pen and paper and tools like Excel. Commenting on storing data, he says that organizations have been obsessed with gathering all sorts of data thinking of the potential.
With the massive amounts of data sitting in storage, Yassin questions if storing all of it was the right choice. He urges organizations to focus on data that can potentially deliver value.
Yassin notes that analytics becomes successful when the data is used. He recommends organizations assess the need for the tool and if it will deliver business value.
Moving forward, Yassin states that defining KPIs can be useful while demonstrating certain expensive initiatives to the C-level for funding. Taking the instance of data quality, he discusses what would happen if a company started measuring data quality before applying it.
For example, Yassin says, having a data quality score of the data based on the signals gathered from running data quality checks could be a way to measure. He maintains that it makes it easier for data owners to do with data governance, as it shows tangible reports on an increase or decrease in data quality.
Furthermore, Yassin states that if the data lineage capability is outside of data governance, the data quality score can assess the quality of data. He adds that organizations can have better analytics and search results by investing in the quality of critical data elements, and as a KPI it shows the value of investment as well.
In conclusion, Yassin states that having a data quality score also helps in maintaining transparency and visibility to consumers. This improves consumption and indirectly improves the entire governance around data and ecosystem, he affirms.
CDO Magazine appreciates Mahmoud Yassin for sharing his insights and data success stories with our global community.
