Data Success – The Easy Way

I hear a lot of excuses for data initiatives failing in organisations, with them saying, “Last time….How will it be different this time?”

“Companies are pouring millions into digital transformation initiatives — but a high percentage of those fail to pay off. That’s because companies put the cart before the horse, focusing on a specific technology (“we need a machine learning strategy!”) rather than doing the hard work of fitting the change into the overall business strategy first.”

Harvard Business Review

This post was inspired by a LinkedIn post talking about buzz words and after listening to sales people promoting their software recently and not giving any airtime to the other elements that contributor to data success. In my view software / technology is only a small contributor to data success. Therefore, I thought I’d share my simple view of how to be successful with data journeys – I’m not going to call them initiatives or programs, as we wouldn’t call finance a program or initiative as everyone knows every company needs finance and they have a role to play continuously in the organisation. So does data. Data has an on-going role to play and does not stop when you’ve delivered a project.

Where to start

This’ll upset a few, but don’t get drawn into software and thinking those in the top right of a recommendation quadrant are the right choice for what you need first.

First you really need to understand what you want to achieve….that’s not a data driven organisation or better decision making. They are outcomes from data success.

Where you need to focus is on where you are losing revenue or opportunities to increase revenue or reduce risk. These are clear value drivers which the investment in data can be measured against.

Examples could be:

• How can we reduce fraud with better data and analytics?
• How can we increase customer retention with better data to personalise their experience?
• How can we reduce time taken to create reports by reducing time spent on correcting data each day/week/month?
• How can our inventory planning be more accurate?
• How can our logistics be better utilised?
• How can we reduce our compliance obligations?

Get the idea, all demonstrable value generating opportunities. I’m sure you can think of many more. If you need more inspiration, check out your business strategy. What has the CEO signed the business up to deliver? How does data and analytics contribute towards achieving each strategic objective?

What’s next?

You’ve got your problem that you want to solve. The success factor is not to get overwhelmed. That’s why choosing one problem to solve will drive success for continued progress. In assessing your first challenge. Think big benefit that can be showcased. Medium effort to resolve. Won’t take months to demonstrate your first success. What can you show in weeks? Then think of a small bounded problem that is not going to need all the departments in the organisation to change for your success.

Consider what you need to solve your chosen problem.

Let’s walk through a couple of examples:

Let’s take ‘How can our logistics be better utilised?’

For example you are either shipping air with part filled lorries/aircraft or you have allocated too much stock for the size of the lorry and it doesn’t all fit on.

Starting to work backwards, checking data as we go. Taking the load the was too big for the lorry.

• We need accurate data for the lorry size and the products
• We need data quality for checking data on the size of lorries and products
• We need data governance for ownership of the data
• We need data quality management to have a process in place for report data issues, remediation, reporting against metrics
• We need technology to drive the ease of managing data with built in AI to develop a consistent approach where insights are revealed to Data Stewards to action or approve

Now let’s try this approach with; “How can we reduce our compliance obligations?”

For example we could be paying a premium as we submit estimated figures, as we don’t have a way of demonstrating otherwise. We need our compliance data used for submission to be auditable.

Working backwards again.

Do we need to amend / request a change to submit reporting based on actual vs estimated data

• We need processes that are documented sufficiently well for auditing
• We need to understand what analytical activities take place on the data, does anything change or is it only a pulling together and categorisation of data
• We need to understand what data is needed
• Is the data categorised correctly, if categorised
• We need to define the checks that will take place on existing data to move it from estimated to actual in the eyes of the compliance authority
• We need to define and schedule regular checks. The process, the selected products, whom is responsible, etc.
• We need to know where the data originates from to understand why we only use estimated data and what happens to the data on its journey throughout the business
• Is any training required to improve the accuracy of data that is collected

Are any process changes required to improve the quality of initial data – it costs between 4 to 10 times more to correct data, than for it to be initially entered correctly – an example of a process change could be that when the first product delivery arrives its quarantined until a full check is carried out to compare actual against system data.

Check back regularly to see if your actions are solving the problem, if not, reassess, adjust and continue your journey.

While you are solving your problem you will no doubt come across areas where you are missing the component needed to carry out the data activities appropriately. Below is an image of my Data Governance Framework, with the components I suggest from my experience are required. Again don’t try to implement all at once. A few key components at a time in a small area to move your problem forward to success one by one is the route to success with Data.

I’ll advise in more detail on each of the components of my Date Governance Framework in a later post.

In summary

Choose carefully, your problem to solve

Work backwards from your problem to determine your route cause and the data needed

Shout about your success

Start small, think big, be successful…thank you, next