1000 days of Data - Reflect and Project

31 August 2019, 3 years of Data transformation, a bit more than 1000 days...

I remember that in 2016 a lot of colleagues, across industries were starting a similar journey. So how can we look back, evaluate and compare in a way that is not anecdotal but - noblesse oblige - underpinned with data? And more important, how can we focus on what needs to be done in 2020 and beyond?

Thinking about this, I made a small self-assessment tool that I would like to share with you. In this article, I'll briefly describe how you can use it. I thought about making an on-line tool, but A. my programming skills are rusty and B. it creates all kinds of confidentiality concerns, so I'll just share an Excel file (will share Excel template by private message upon your simple request) with fictitious sample data. Up to you to put in your own data and who knows, maybe you want to share the results. You decide of course.

I used 4 moments in time, 10 content dimensions to reflect upon and 5 levels of scoring.

The 4 moments in time

For me, it was relevant to go back to my starting point, 3 years ago. Next, I measured the as-is of today. Third, I looked forward at the ambitions for end 2020, short term. And finally, I plotted the ambition for another 1000 days down the road.

5 levels of scoring

Scoring is tricky, so I keep it simple. Score 1 means that your organisation was NOT doing anything at all. Score 2 means that the activity exists, but rather in a pioneering, non-orchestrated mode. Score 3 means that your organisation has consciously organised the activity, making it part of the formal project or organisational governance. Score 4 tells me that not only are you organised, you are showing tangible benefits. Benefits that are recognised by the owners of the P&L, not just by data-enthousiasts. Score 5 is ambitious. 5 means that your activity clearly contributes to the competitive differentiation of your organisation, that it's visible to the market and external stakeholders, or (deliberately) protected internally (e.g. a specific algorithm). The scoring below 4 is effort oriented, as of 4 it is clearly impact oriented.

10 content dimensions

You can write books about every single of these dimensions. So you will need to transpose their content a bit to your specific company and industry context. Remember, this is an instrument to help you, not to fool yourself.

Dimension 1: General Data Governance and Organisation

This is the extent to which you have succeeded in embedding the elements of Data Governance that are relevant for you inside the existing organisational governance bodies. If you can re-use existing bodies and embed data into them, that's the most effective. Does your organisation have teams (including "teams-of-1") that wake up in the morning for data definitions, data dictionaries, technical and business meta-data definition, data models, etc. Do you have a meta-data management system (I guess your CFO has an accounting system for the money of the company, so you should have yours too for the company's data)? Are there agenda topics in your management meetings that zoom in on these aspects? Do these elements pop-up when discussing investment priorities? Does your HR recognise specific data functions, just like it does for commercial or financial or production functions? Do you share best practices with others, with your regulators?

Dimension 2: Data Quality Governance and Organisation

Data quality governance could be considered to be a subtopic of the previous dimension. However, it is so critically vital that I make it a separate dimension. Have you defined which critical data elements need permanent quality monitoring (imagine a chemical plant without sensors)? Have you reduced the operational cost of quality monitoring by automating it with a quality measurement tool (imagine the same plants with people measuring the temperatures all over the place manually)? Do you have a process or tool to do Data Quality Incident management and remediation? Do you report on data quality evolution in your management dashboards? Do you have a regulator or other stakeholders who monitor this? Do you have specific HR-roles? Is your client befitting from better (client?) data quality?

Dimension 3: Analytics A.I. Governance & Practice Maturity

Data science, analytics, A.I.,...no matter what name you give to the beast. Let's call it A&AI for simplicity. If ever you want A&AI to get out of your labs, you will need to organise and govern it decently, just like you do with other important IT deliveries. Do you have data scientists and did you make a deliberate choice about how they are organised? Everybody has a different organisation, but did you think about yours? Or did your organisation "just happen"? Do your people structurally exchange knowledge? Are their toolsets aligned with IT tools and standards? Do they check upfront with your management whether their ideas would withstand public scrutiny? Do they run their project systematically through a legal and compliance check? Do you know who will look at the code once it is running every day in production? Did you agree on who will check the validity of the models on a recurrent basis? Are you decommissioning the old technologies or are you just adding to the long term technical debt with today's innovations (which are tomorrow's legacy).

Dimension 4: Analytics & A.I. Capability & Impact

Do all your business stakeholders know what A&AI can do (really, not the sci-fi version)? Are all new technology projects and new business processes challenged so that you detect in a systematic way how these new techniques can contribute? Are you merging A&AI into traditional technologies (e.g. scoring) and processes (e.g. helpdesk) or in other new ones (e.g. augmented robotic processes automation)? Are you discussing how to move from rule-based approaches to probability-based approaches? Are your regulators involved? What are your external stakeholders getting from all this (artificial or not) intelligence?

Dimension 5: Business Intelligence Capability & Impact

Ah! Business Intelligence. Many times proclaimed dead but still alive and kicking! But have you been rationalising your data warehouses? Have you been keeping them in line with new requirements and processes? Have you been working on documentation, code automation, data visualisation,...? Have you been killing all the dead reports in your organisation? Have you managed the diversity of end-user technologies? Are the business users seamlessly integrated in your prioritisation (agile-way-of-working?) and governance? Have you dealt with the costly proliferation of "management-by-manual-powerpoint-reporting"? Did you modernise the career tracks of your BI-specialists with perspectives towards big data, A&AI,...? Are you improving your reports to make sales or operations more effective? Are your external stakeholders seeing any of this?

Dimension 6: Data Visualisation Capability & Impact

Could be a sub-dimension of Business Intelligence, but due to the relevance, I gave it a separate spot. Have you invested into understanding which visualisations work best for which type of messaging or decision taking? Did you make a clear technological choice and are you developing a collective expertise around it, allowing also for self-service capability? Are you actively decommissioning traditional reports and replacing them with (fewer) visualisations? Are you connecting the "decisional-powerpoints" directly to live data sources? Are your external stakeholders benefiting from your visual insights? Are your clients giving you better scores because they love the way you visualise their (yes, it's theirs) data?

Dimension 7: Analytics & AI Infrastructure Maturity

Is your A&AI infrastructure at a level of "production"? Is it backed-up, documented in your ITIL or other continuity procedures? Do you know which libraries your data scientists are using and are they security and stability approved? Can you generate synthetic data to be more efficient with regard to data privacy concerns? Do you really need GPU's - and if you do, do you have them? You're doing great things on the cloud, but can you go to production in the cloud? Is it integrated with your other technologies (data flows, security, ...)?

Dimension 8: Big Data Streaming, Storage and Processing Maturity

We all talk(ed) about Big Data (high volume, high variety and high velocity data). But do you really have the infrastructure and skills? Your infrastructure can be cloud based or private. How are you dealing with the massive fragmentation of technologies and skills in the market? Are you consciously building your own knowledge? Do your big data projects make it to production? Can your client more enjoy your services because you moved into the big data technologies? Does your traditional, role-based security model still hold or do you need data-based security?

Dimension 9: Care for Data Privacy and Ethics Awareness

Sure, you are GDPR compliant. Of course you are. Did you think beyond GDPR? Did you decide what your moral/ethical compass is? Do you adjust your client facing communication regularly? Do you have ways to detect what IT or data scientists are working on? Do you have day-to-day processes that support "privacy-by-design"? Are you removing old data, every day? Do you still have budgets to do this? Do you report about this to your internal and external stakeholders? Are you in the public debate?

Dimension 10: Data Monetisation

What can help you to achieve scores 4 and 5, to make an impact, is to think about Data Monetisation. I do not tend to do product placement, but if you want to structurally think about this, the most efficient way is to read the book on Infonomics by Doug Laney. Do you talk about data as an asset that has a value? Do you know from whom you buy data? Have you identified for whom your data or your insights might have a value? Is your top management aware of the new data-driven business models that are popping up? Are you sharing benefits of data monetisation with your clients (it's their data in the first place)?

I have a score, so what?

I use this approach for the following purposes:

• Give people a sense of achievement and perspective. People tend to not always see the big evolutions to which they contribute. Going three years back and looking at things from a high level perspective, gives a sense of achievement and pride.
• Force yourself to self-reflect and think about impact. It's great to do "a lot of stuff", but you need to get to levels 4 and 5 to be sustainable. Just doing "data" because you believe in it, will not keep your team going, nor will it secure continued investments.
• Communicate with stakeholders. Data-people sometimes have their own jargon and are often "believers" in what they do. But at the end of the day, you need to reckon with general managers, shareholders, investors,... And having KPI's and showing that you care about them really helps to communicate crisply in a format that they are used to.
• Makes you think. Makes you project. It makes you think about achieving levels 4 and 5. It makes you think about the future. From thinking, will come talking, inspiration seeking, ideation, creation, future making.

Good luck, I hope it's relevant for you.

Jo