Data Management is one of the most trendy topics in recent months. Many data initiatives have been developed without good foundations or structures to support them, a situation also caused by the silos that have been naturally generated in organisations. However, companies are beginning to realise how important it is to establish a good foundation on which to build a data-driven organisation. Some companies are taking action at an early stage, and others have had to put the brakes on in order to restructure in a more organised, governed and sustainable way. Thus, Data Management has become a priority for companies looking to evolve as a successful data-driven organisation.
At Keepler we have dealt with organisations at different stages over the years, especially since we created the specific area of Data Management Consultancy in 2022, which has allowed us to get to know in depth many real cases of data management that organisations are experiencing.
This experience has led us to identify these 7 trends in the field of Data Management that we foresee should be very much taken into account in 2023.
1. Operationalising the governance model
At Keepler we continue to encounter many clients who, with a defined governance model (on paper, and generally very ‘standard’) are not able to put it into practice or to get ‘data accountability’ working in the day-to-day life of organisations. And it looks set to remain a very common scenario in 2023.
Part of WHY this happens is a consequence of using the typical approach of operationalising Data Governance through tools, which are not usually free (at least those with a suite designed for functional users, which are 95% of those involved in a data governance programme) and whose investment justification is difficult to sustain (precisely because they do not understand the ‘return’ of the data governance programme in directive layers).
Another common case is BECAUSE the governance model defined on paper does not adapt to the reality of the company (due to a lack of profiles with certain technical-functional skills, organisational culture, heterogeneity of business units, etc.). Defining a customised data governance model for each organisation and being aware that a Wikipedia model is not necessarily the best solution is part of the beginning of this common pain point.
2. Redesigning the Data Office
Although we thought this was going to be a rare occurrence, throughout 2022 we have come across several organisations that are reviewing the best organisational model for the data function.
Many companies are still starting to design a Data organisational cell as a spin-off from their digital teams (more commonly) or even still from technology (data under the direction of the CIO has little scope). And it is in this situation that doubts arise about both the suitability of the location and the separation of powers (functions) of the new Data area to be sized.
Organisations are also beginning to appear which, with a clearly mature Data Office, are considering needs not so much in terms of organisational redesign but above all in terms of talent and development of the profiles of the Data team. It is very interesting to see how the stewardship models as they were conceived fail (they have no long-term development), although this is a trend that we believe will be contrasted and extended in the future, it is still very incipient.
3. Data quality remains an unresolved issue
In few organisations have we yet seen a strategic use of data quality (with meaning, with the intention of driving change and achieving cross-cutting objectives). And it is undoubtedly the best weapon a Data Office has to bring about a change of mindset towards data in its organisation.
Data quality has to be used as the main tool to empower those involved in the Data Governance Programme to perform their functions with visibility of their impact, and it is not always done this way. It must be.
It is also necessary to give an understanding to non-technical users (who represent sometimes more than 50% of the total data quality efforts/needs) about what data quality is or how it is calculated. Something so simple is not being done and is starting to attract more and more interest from CDOs and Data & Analytics leaders.
4. Cloud platforms extend catalogue capabilities into data governance
In recent years we have seen how the main public cloud vendors (AWS, Azure and GCP) have been evolving their data catalogue solutions towards more and more front-end and functional user-oriented functionalities.
In this sense, we saw how Azure’s Purview was positioned somewhat further ahead of its GLUE Data Catalog counterparts from AWS or Google’s GDC, but with a more ‘defensive’ approach oriented towards regulatory compliance and control of data assets from a hub. At the last AWS re:Invent we have seen DataZone (with a more ‘offensive’ and self-service oriented approach) presented as a solution beyond the old GLUE Data Catalog and the outlook is that this is the way forward. On the other hand, GDC has integrated its GDC catalogue service into another service with data marketplace ambitions called Dataplex.
Although we do not expect that in 2023 the large public cloud vendors will be able to raise their current tools to the level offered by traditional vendors of Data Governance solutions, we do believe that they will continue to invest in this line and get closer and closer to the end user (mostly non-technical). In the coming months, they will seek to cover and develop functionalities that were not contemplated until now (workflows between roles in the governance model, alerts, data marketplace, glossary of terms, improvements in lineage visualisation, etc.) on top of their current capabilities of complete integration of the metadata catalogue and lineage.
5. Engaging the organisation in data & analytics initiatives
This is a consequence of both the difficulty in implementing data governance and the lack of strategic use of data quality.
Many companies have been deploying communication and change management programmes for years, some even have Data Literacy programs in place, but the “last mile” is still seen as a failure.
In this sense, we see that work will continue and we believe that in 2023 this aspect will continue to improve a lot, perhaps it will be time to see great achievements and milestones in terms of change management and the achievement of a mindset that is no longer just data-driven but data-asset (making it clear that data is a valuable asset and not “something from IT”).
6. The great unknown that everyone is talking about: the democratisation of data.
A very recurrent topic of conversation but one that is rarely translated into actionable and simple to understand and address. Companies still want to be data-driven, achieve self-service and self-consumption and be very mature in terms of data democratisation. Yes, but what does this translate into? There is silence.
We will keep on talking about the data democratisation and there will be an increasing interest in knowing how to calculate it, something we have not yet seen and for which Keepler has developed our own calculation formula: the Data Democratization Index (DDI), a practical and grounded way of measuring the degree of democratisation of real data, which can be tracked and monitored to measure the impact of our data & analytics initiatives and their reflection in the daily lives of users or not.
7. The jewel in the crown: the value of data
Undoubtedly, this is, will be and will continue to be the CDO’s great challenge for several years to come. Knowing how to measure the value of data (both intrinsic and derived) and the impact generated in business terms by the different data & analytics initiatives led by the Data Office. Even some ambitious or challenging CDOs like the idea of building themselves a P&L, and it seems that this will become more and more common.
In our vision, we advise our clients to start step by step, that is, to begin by measuring the impact, translated into business terms (operational efficiency, additional revenue and risk mitigation), of data & analytics initiatives. A measurement agreed with the business (or functional users involved) through metrics previously established and contrasted with measurements prior to the implementation of these D&A initiatives.
Once we have measured dozens of D&A initiatives and learned to refine assumptions and previous premises in order to know more and better how to estimate the future impact, we will be able to move on to a second phase of measuring also the intangible impact (culture, customer or even employee experience) or even the intrinsic value of the data (the monetisation of data that is so much talked about and so difficult to achieve, similar to the democratisation we mentioned earlier).
This is a long way to go, haste was never a good advisor and our recommendation is to start walking before learning to run fast.
Data Management Consultancy Lead at Keepler, "I do enjoy working with big organizations and help them understand and truly believe how better data management capabilities mean better business results."