The Data Leader Do’s and Don’ts During a Crisis

Ido Biger
Data
June 27, 2020

https://images.app.goo.gl/86TVGs2gTjzZ7RX88

The Covid-19 pandemic took us all by surprise, causing many to be frustrated, unemployed, on unpaid leave and forced to work remotely in a situation where every task has always been carried out traditionally. The pandemic has disrupted the economy of virtually all the countries affected. However, in the moment of crisis, there is always a ray of hope for everyone in diverse ways as it occurs to them, it is left to us to hold on tight or let it fade away.

This article will focus on the practical side of this crisis, most especially to the Data Leader, the one who lives and breathes numbers and acts as either the reflector of the actual situation or even more importantly – the differentiator as far as pointing out the right path of overcoming this situation in the most relevant way.

Firstly, I’d like to say, this article is based on my experience and that of some of my colleagues.

So let’s start with what should we do…

1. Understand the business impact – focus on the financial / accounting terms.

As always, the data leader should be a native speaker of his/her organization’s language, nevertheless during crisis one should fully understand the impact in each of the units in order to be able to reflect it (through reports/dashboards/notifications) and more importantly – make a wise assumption on the upcoming challenges, by developing the most relevant prediction models and forecasting methods.

As for improving our financial abilities and understanding, there isn’t a better and more important time for data leaders to fully understand the accounting terms. AP (Accounts Payable), AR (Accounts Receivable), Accrued expense, book value (BV), COGS, Dep, GM (Gross Margin), CF (Cash Flow), Liquidity, are all terms we data guys should add to our arsenal, not only to speak the language with our Finance & Accounting data customers but in order to be an integrated part of the solution/plan (they’ll come up with). It was only when I’ve drilled in the company's Cash Flow model that allowed me to fully appreciate the complexity of the elements and confront it technically when developing the CF data product.

2. Be part of your delivery line – don't lose your hands on capabilities!

It is a known fact, that once the data leader runs and manages a large amount of employees and especially managers in different hierarchy levels, the data leader allows him/her self to focus more on the data strategy and over viewing the data product line lesser than the actual delivery line and the ground floor fundamentals. Of course, there isn’t any expectation for the data leader to be the best data engineer, the most skilled data analysts or the most brilliant data scientist of the team, but one should not forget that his/her hands on capabilities are the ground rocks of the leader's knowledge. managing is a very important part of his role, but especially during crisis, it should not be the only characteristic one withholds. It will never replace the need to understand, to learn and to be able to actually be part of the delivery line. I’ll add to that a few important points.

The front end deliverables, whether it’s the BI products (dashboards, data models) or analytical tools, are the most significant know hows (as far as hands on) for the data officer. It is what allows him to fulfill all the urgent needs as far as business question even if the data level isn’t well structured. By well defining the semantic layer, one can come out with great data products even if behind every object relies a variety of systems that are not best structured for the solution. Which leads me to the rapid bi methodology.  

3. Rapid Data development – always, but especially now.

One of the biggest known mistakes we’ve done over the years as IT personnel and technology geeks is to fulfill our technological dreams while trying to deliver our customers wish-list. I urge my employees to turn it upside down. The most important part of a business need is making sure we’re putting most of our effort on the most valuable part of the product. As far as data products – it would be the outcome – the deliverable itself – whether it is the data model of the report. Therefore we should start with the finish line – the deliverable.

No alt text provided for this image

The Blue BI product retrieve its data from the DWH, a website, the BD platform, a certain cloud service and one operational system. instead of going the traditional way, by analyzing all the sources and structure them into a finalized model in the DWH, we're focusing on the delivery itself, and supporting its needs in the semantic layer (preferably with views). the collection layer will get the exact necessary outputs (integration tool or DB-links).


The same methodology applies with the orange Rapid BI Product (with its relevant sources and requirements)  

The Rapid BI methodology suggests that the business customer should get his deliverable within days. The architecture behind can be understood easily in the diagram above.

The constant argue with the BI Manager (in my case - reports to me which makes it a lot easier than in cases where the BI manager doesn't report to the Chief Data Officer) would lead to a thriving environment in which the products keeps coming and being valuable for our customers while making sure there is a known list of data engineering projects that are being deployed with a well-defined end game (as seen in the diagram below) - the product is already in production.

No alt text provided for this image

The Blue BI product retrieves its data from the DWH (!) which collects it from a website, the BD platform, a certain cloud service and one operational system.

since the data product was already deployed successfully (the back and forth with the customer was developed previously while the rapid product was defined)


The same methodology applies with the orange Data Product (with its relevant sources and requirements)

That allows the organization to be fully active with relevant data products, up to date and significant alongside making sure the data engineering team is focusing on the relevant products with a minimum waist of effort.

During a crisis, when your team decreases to the minimum possible, the delivery line should be utilized in the most efficient way, thus, starting with the end products in the quickest way would ensure we’re focusing on the most precise deliverables. Each change / fault would be fixed very fast due to the fact that it is on the surface and not relying on several layers of integration tools.

4. Real Time Data Models

As expected, the focus during a crisis would be towards here and now. looking months away, analyzing years back, it is all important and will still be the fundamentals of the planning process. And yet, the main differentiation between response and act would be the ability to supply up to date information. by the hour, by the minute. it will allow the decision makers and business owners overcome challenges as they emerge (and relatively still small).

When choosing where to start with the real time models - I would suggest the following criteria -

A. the most relevant technically (feasibility, first make sure the data is available for your team - whether via integration tool or other means)

B.the business owner - if the decisions made by the BO/system are rapid and swift (such as digital systems that responds to an event), one might start there rather than focusing our efforts on the steady ones.

5. Keep your team (those on unpaid leave) informed but not involved. Utilize the medium of online training.

It is a crazy time; this period is frustrating as ever, people are requested to stay at home, mostly without payment, and in the worst case, firing them with nothing to rely on to get them through the period of isolation. It is important to keep them informed and updated at least once a week, on the progress and challenges. But trying to keep them involved with the delivery line and tasks while on unpaid leave, even asking questions and looking for guidance would increase the frustration and would create another crisis in the middle of a crisis.

We’d want our teams to come back more appreciative, thankful and of course energetic towards the uprising / recovering period of the company. They should be allowed to be free from tasks and at the same time focus their efforts on learning online for the purpose of improving while also being with their families.

Ido Biger

Ido Biger is the Chief Data Officer of ElAl Airlines. Israel's National Airline.In charge of making ElAl a leading Data Driven Organization. Managing the Head of BI, the Data Science team and the Data literacy program. Reports to the company's Chief Digital & Information Officer.

Before joining ElAl, Ido was the Chief Data Officer of yes Television, where he was Managing over 30 Data engineers alongside being responsible of the Business Analytical units and the Data Science team in the company,Reported to the CIO as the Head of BI & Data and to the CMO as the Chief Data Officer. International Speaker, Big Data Technologies and Data Visualization adjunct Lecturer at Tel-Aviv University's MBA and data science business practicum adjunct lecturer at the Technion – Israel Institute of Technology.

Related Posts

Newsletter BrazilClouds.com

Thank you! Your submission has been received!

Oops! Something went wrong while submitting the form