Understanding the elements of data-driven decision making - Business Analytics

At the beginning of the outbreak of COVID19, few predicted its global impact. This was assumed and shrugged off as a viral confined to China. It was nobody's problem till reality dawned in March 2020.

We failed to foresee due to inherent bias..!!

This happens when business managers don't want to add new variables to their existing portfolio of problems. We tend to believe that the worst will not happen to us and remain optimistically ignorant.

Business Analytics not only brings a raw picture of your business but also forces you to think in multiple directions, an inconceivable task by managers already multitasking. 

In the current situation post the outbreak when businesses are struggling to find the ground, Business Analytics is the key to get out of the woods or at least know that if it's even possible in the first place.

Business Analytics can help managers to rebuild their strategy with higher efficiency and accuracy. It can bring the required confidence in steering the energy in the most optimal direction so business puts its best foot forward.

How can Business Analytics help us during a crisis or otherwise as well?

It's a subjective topic and answers will differ with changing context. We can look at a generalized approach that is business-wise authentic, mathematically sound, and substantially future looking. The steps can be illustrated as below. Sometimes the order may change a bit but the essence remains.

Identify your moving parts and levers

Quantify the interactions

Build a singular source of truth

Create a desired future state

Design a "What If" framework and iterate through it to reach the desired future state

Channelize resources to activate the simulated value of the levers

Revisit the framework with new inputs to alter strategy

The steps are explained below with easy to understand examples.

Identify your moving parts and levers

Managers have to identify all the participants in their organizations and businesses that may impact their KPIs. Partners, raw material, employees, countries' unemployment rate, number of rainy days in a year, and so on.

Every business be it small SME or large scale enterprise or a LOB has multiple dynamic variables. One can count them as, Input Uncertainties, Processing Issues, and Output Challenges.

Input Uncertainties springs around your input variable like fluctuation in the supply of raw material, change in the availability of resources (movable and immovable), There could be external variables like regulatory, legal, environmental, and so on beyond one's control. COVID19 being one. Outputs Challenges are your impediments in producing your final product, your value creations (tangible or intangible). Everything remaining can be put under Processing Issues which are within the ambit of your organization.

Any of these moving parts are correlated with the performance of one or more individual or organizational sub-units in the organization. The task is to identify these sets of people or business functions which may be impacted if the moving parts move too much. 

E.g., if there is a significant shift in the supply of raw material, the supply chain partners get under pressure as their incentives may be linked with the slabs of their supply volume. It inturns impacts the capacity of products to be produced stressing your assembly line staff. And finally impacts your capacity to produce the output or the volume of business you can do, and eventually disturbing your market share and top and bottom lines. It's a dominoes effect corresponding to a change in a moving part relegating to your final commitment.

The key is to identify all the participants in the game of business who may directly or indirectly impact your capacity to create value.

The ones you can control or manipulate will be your levers and others will be your controllers.

You can play with your levers only to an extent this can be called an "allowed threshold". It's important to keep an eye on the controllers while keep moving the levers to your advantage within the allowed thresholds.

The Business part of the business analytics can help identify these stakeholders but mostly this comes from a joint exercise between multiple department heads coming together and contributing. The Business Analytics consultant having the capacity to interact with different stakeholders seamlessly becomes significant.

Quantify the interactions

The moving pieces as elaborated in the example above interact with each other in one way or the other. The key is to identify how much is the influence of one on another? i.e. Once you have identified the interactions, you need to quantify those.

E.g. If employees are absent by more than x% then how much will it impact my production line or my sales targets? What is the threshold under which business can sustain its desired pace of growth? If impacted how much the business may get impacted? While a lot of it requires business acumen and expertise from your seasoned managers, the rest goes into an analytical model to produce numerical quantification of the interactions among the moving parts. 

Business Analytics is a blend of mathematics with historical data and some business assumptions to predict future of the business!

The key is to link numerically how your favorite KPIs are impacted if the variables shifts from their place. Business Analytics can help you get your numbers right.

Create a singular source of truth

Businesses need to harvest all their available data to create a unified flow of information following 4 Cs

i.e. Current, Consistent, Comprehensive and Consumable

Current

The data need to be as realtime as possible. This can be achieved by automating all the data channels, create data pipelines that can funnel the data seamlessly to a singular place or a data lake. In smaller businesses, this can be a humble cloud-based database automatically fed by spreadsheets arriving from different data gathering employees or data generating software. Smaller businesses often make a mistake by assuming maintaining a database is not for their size but that's not true. Cloud-based databases are affordable, resizable, secure, and accessible with most of the visualization tools. So a must-have.

Consistent

All the road should originate from a single source of truth, which means the data should be accurate and logical from one report to another even if you have multiple sources of data. The soul should be common. Although it's advisable to maintain a data warehouse that can house all the data keeping its integrity intact. But often in businesses, it's not possible as data resides in all shapes and sizes from spreadsheets to transactional data in a database. Data integrity is a key virtue of the soul of decision making.

Comprehensive

Everything needed to make business decisions must be recorded with accuracy, stored with reliability, and aggregated with the confidence of being whole in itself. It's an ideal scenario but if one can restrict as many loose ends as possible from framing a complete story, amazing decision making is plausible.

Consumable

The reports and dashboards should represent a harmonious dialogue with the user's problem with minimal room for misinterpretations. The reports should be aimed at solving the problem of the reader than just spitting out data in charts. Not to mention the hygiene is to be maintained at all times like charts being neatly labeled, positioned in a logical order with correct placement of filters, and so on. A comprehensive guide on how to do reporting is out of the scope of this article. The bottom line is, there should be value in every pixel of the report.

Business Analytics can consult with data engineers to sanitize the data as per the business rules, bring them to a data platform via data pipelines connecting varied sources of data collating current information consistently, creating a comprehensive version of truth for your organization which can be consumed uniformly.

Create a desired future state

Once you are there with your building blocks, create the desired state of KPIs aligning with your short and medium-term goals. The goals may not look feasible at first but when it's based on a quantified analysis then it all starts to make sense or at least you will know if it does make any sense or none at all to proceed with.

E.g. We should contain the risk of churn of our depleting customers within X% +/- 5% while Y% can be labeled as reliable and nonrisky. 

Or 

We need to keep our profits to X% without dis-engaging Y% of the clients.

All the KPIs mentioned above should already be part of your source of truth exercise. Meaning you should be able to quantify them and be able to measure the impact given the moving pieces find new places.

Business Analytics cannot spell your goals. This is top driven activity and only acts as guidelines for a business analytics team.

Design a "What If" framework and Iterate through it to reach the desired future state

By now you have got your nuts and bolts along with your desired destination. Time to experiment with a predictive/prescriptive dashboard or a framework which can help you simulate various business situation while carefully altering your levers or other controllers. This result would be enabling yourself with multiple combinations of possibilities to reach your desired future state or at least get confidence that the same can/cannot be achieved. If not then by how much? Or What additional needs to be done to scale up to the task?

One can imagine it like a simple game where A influences B by a factor of X and B influences C by a factor of Y. Then how much A influences C? This way when 100s of parameters participate at organization, department, or a product level then a combination of maths and predictive science can help you navigate the complex labyrinth to map your goals.

Key questions to be asked to yourself

Am I able to simulate my variables and see their impact on my KPIs?

Are all the variables covered in the simulation?

Have I outlined all the interactions?

Do I understand the limitations of the extent of my variables?

How accurate are the relations defined among the variables?

Does the framework help me visualize a path to the desired future state?

These frameworks can be as humble as an excel model with few variables and equation linked outputs, typically known as a business case for a given output, all the way to a sophisticated self serve dashboard or a data cube with an ability to query yourself. More the sophistication better handling of the organizational complexity but takes more time to build and test. Most of the time one can start small and keep building on top while iteratively testing the existing framework.

Business Analytics can help you build your tools to experiment with multiple business scenarios and outcomes, will also inform you if the same is achievable or not, and if so by how much or how less!

Channelize resources to activate the simulated value of the levers

Once all the possible scenarios have been visited, iterated, agreed and approved, its time to channelize your energy to move the needle aligned with the strategy framed above. The key for a leader here is to understand which area needs more focus than others. Again the framework(s) build above can be used as guide rails along with the understanding of the dynamics of your department. Another dashboard with the KPIs on employee productivity can be useful to measure the success of your initiatives in a timebound manner. 

Business Analytics reporting on every aspect of your business can help you keep track of your changing business performance.

Revisit the framework with new inputs to alter strategy

Well, we are not living in a simulated world, this is real. So now you will get feedback from the field which you may not have anticipated. There will be news of your assumptions going wrong while the reality was something else. Under all such circumstances, you need to revisit the framework with fresh eyes and incorporate all the feedback. Perform the simulation again with the same focus and realign your resources. The more agile the better and faster the outcome.

Any manager working within the above framework will reduce the uncertainty around decision making in his business unit and proceed with utmost optimization and hence likely to attain success more than going about gut-based random decision making. Not that it cannot be right but it does need all the parameters to fall into place by chance.

Are you willing to depend on time to decide your fate or are you the one to take control of the outcome of your decisions?

By Abhinav Kumar

Reachable @ https://www.dhirubhai.net/in/iamabhinavkumar/