GENCFO Clubhouse
”Community via conversations and networking”. Following the demise of ‘Clubhouse’ we have started our... View more
Host:
- Organized by
-
-
Data management in Finance
-
Data management in Finance
It’s been really interesting hearing the discussions here over the past couple of days and there are many themes that I recognise. I thought it might start an interesting conversation to share some of my experiences as a data geek in a Finance Department.
I work in the Analytics and MI Team within my organisation which is part of our Treasury Department. My background is data, and prior to this role I worked in the Health Department as a Hospital Informatics Manager.
In my current role I work closely with Finance Business Partners and some Budget Holders across the organisation. We have lots of Excel use here and there are many regular tasks that start and end with Excel each month (taking up massive amounts of time for our 50+ FBP staff).
We noticed that many of our FBPs are essentially performing the same tasks in Excel each month with a different file, and filtered views of the same data. Much of their time was spent on report production, less was spent on the identification or sharing of actionable insight or decision support.
Now this hurts my lazy mind. The automation opportunities were huge.
After some more digging in to the issues, I realised that we could help by improving our data management processes.
The key principles that we started with were:
– Control our data inputs – direct from system where possible, structured, validated input where not possible.
– Transform the data once – fixing data formats etc.
– Store transformed data in a data platform – SQL server was what we had available to us
– Assess and flag data quality/integrity once – notifying data owners of issues to fix at source (not in our platform!)
– Report from and process from, the single version of the truth – this allows full access control and sharing
– Report what is in the platform – if it is wrong, fix in system or source
With these principles in place, we were able to build a platform that updates daily rather than monthly, allows us to automate all report production and frees up a significant amount of time across the department. Access to data is easier and more consistent and everyone starts with the same thing.
My data background also led me to consider the governance aspect of this data and platform so we built many controls in order to comply.
So that sounds great right? Sort of…
My biggest learnings from this work is that by taking this on, outside of our local IT (they had no resource to support and an enterprise solution is a while away), we miss out on a lot of the really good practice around data engineering and development. These are the bits that ‘keep the lights on’ and are business as usual within IT.
So while you can build yourself a tactical improvement like we have, to truly achieve sustainable transformation, you’ll need IT, Data and Finance Teams to work together. Taking shortcuts needs to be planned carefully to ensure you don’t develop technical debt that ends up causing problems down the line. We worked with our IT and Central Data Teams to ensure that what we were building aligned to our longer term strategic objectives.
My plan for the future is to take this to the next level and I am changing roles next year to help drive the enterprise solution as an Information Architect!
How are others engaging with strategic developments around data? Do you even have any yet? How does the IT/Business relationship work in your organisations with regard to data?
-
10 Replies
-
@SamL can we be mates? ????????
-
???? sure!
-
-
Excel-lent Sam! See what I did there….
It is so true and why GENCFO exists, to learn, and to grow…
I have never understood why anyone would invest so much into their career to “performing the same tasks in Excel each month with a different file and filtered views of the same data”…
You nailed it! It’s time for change!
Do spread the word at GOV, as it’s so important to get these messages out there!
-
Very good Chris!
-
-
I once got fired from a short term contract for setting something like this up – daily performance reporting vs forecast based on automated nightly exports from an ops booking system. The local Ops Management *loved* it – life changing for them. The Chairman fired me. If I properly understood why, I’d tell you. Maybe this is why I have stakeholder management issues @mrmikerose ????
-
Wow… sounds like you were better out of there Lynne!
-
-
-
More than happy to chat about them (I have helped put together and am now looking at using in practical applications… including the 2 private beta tools linked at the bottom of this site) -> https://theodi.org/news-and-events/news/the-odi-announces-new-framework-of-data-practices-and-private-beta-of-two-new-tools/
theodi.org
The ODI announces new framework of data practices and private beta of two new tools
The ODI announces its new framework of data practices and a private beta of two new tools to help organisations assess whether their data practices are trustworthy at the ODI Summit.
-
Thanks @chris-a and @mrmikerose that looks great. Might give the pilots a go.
Internally I’m pushing hard for a standardised data maturity assessment. One of my favourite quotes from one of the models is:
“It’s common at this stage (level 2 of 4) for the organisation to demand innovation beyond data capabilities or capacity”
This is a really important point IMO. As member of a data team, we are often asked for the shiny new things that are seen in media etc. (think of “AI”).
The reality is that without your solid foundations of dataflows, governance, data quality, sharing etc. At best, your shiny things will be useless. At worst, very damaging.
The data maturity models are great, as you can build a timeline to improve maturity score and flag on the timeline when you will be mature enough for advanced outputs. You can also assess department by department and introduce a common understanding of a simple score.
We’re using this to help prioritise the ‘boring’, under the surface foundational work that needs to happen first. No-one gets excited about version control on data pipeline code but it needs to happen!
-
Agreed – data maturity models can be very helpful.
Although, sometimes, they are not accessible to non-data people… which is where a skill and experience in how to communicate what a maturity assessment actually means can be important 😉
-
-
Log in to reply.