For business leaders seeking to future-proof their organisations, data should be the foundation for all future-proofing strategies. How can small businesses use data to predict their future performance?
Given recent global events, businesses were reminded just how quickly external events could affect bottom lines. One of the most high profile of these was Toyota, which lost $1.2bn in value thanks to a critical Trump tweet.
But external events can affect small businesses even more acutely, and being adequately prepared for all circumstances can be the difference between boom and bust.
Thanks to the plethora of data that all businesses now create, predicting future performance and pitfalls is easier than ever before. For business leaders seeking to future-proof their organisations, data should be the foundation for all future-proofing strategies.
The first step to any kind of performance prediction is to know exactly what your business’s performance looks like. This means looking at all the data your business produces over the course of a normal working day.
Tracking this data over time – say months or even years, can help you identify patterns in performance, and iron out any possible anomalies.
Some of the data you produce on a daily basis includes sales, marketing, HR, financial and product/service data. In other words, the number of clients you have, the number of leads and sales, how many marketing campaigns you’re running and where, the response to these campaigns, how many staff you have, how many sick days your staff have had, how many have left, stock levels, cash flow and other resources – the list is seemingly endless.
All data matters
Collecting all of this data can be a little tricky, especially if it’s siloed away in different areas of the business. In the course of analysing all your data, you should look at breaking down the silos between your data and putting it all in a central reservoir called a data lake.
Indeed, it’ll soon be a legal requirement under the General Data Protection Legislation (GDPR) to know exactly what personal data you have stored and where, and ensure it’s easily accessible to relevant stakeholders.
So putting the legwork in now, by ensuring all data is kept within a secure data lake, will pay dividends come May next year, when the legislation will be enforced with a €20m fine.
After you’ve collected all your data and formatted it so it can be compared easily, you can then set about analysing your data.
Effects and analysis
Looking at your data month on month, week on week or even better, in real-time can tell you whether your sales, marketing and other functions are affected by seasonal variations, marketing campaigns you’ve run or events you’ve attended.
This starts to paint a clear picture of which external circumstances can affect your business’ performance. If you want to get more advanced, you can use data science to overlay external data like open weather data, infrastructure data, governmental data and transport data on your business’ data to see if each of these has any impact on your normal working day.
All of this data can then be used to build a data science model that represents what normal looks like for your business. This is simply a mix of different algorithms fuelled by your data, that can be used to test different external scenarios.
Test your model
You can use to model to see what could happen to your business in real life if say, there was a new entrant to your market, or two competitors merged to take the majority of market share, or even if there was an earthquake disrupting your supply chain.
In short, you can run the model through all sorts of tests to see whether your business would buckle under the pressure – even or unpredictable and unprecedented external events.
You can also put the model through longer term changes t o see how this gets worse over time, like a chronic undersupply of a specific material your business needs, or a lack of skilled workers.
The model can also be used to test solutions for your business. Say you’ve identified a new market entrant as a potential problem for your company, several options you may have include diversifying your product offering or expanding into different markets. You can run a simulation on your model to see what solution is most effective and least risk.
When forecasting your business’ future performance, you no longer have to rely on gut instinct or past performance alone.
Supplementing your internal data with external open data can help you spot patterns you would have otherwise missed, uncovering external influences on your business you never even knew about.
It may require a bit of legwork in the beginning, but creating yourself a decent data infrastructure and data science models that you can use time and time again can really up your prediction game and drastically reduce your business risks.
About the author
Yariella Coello is the head of consultancy at data science and marketing services company Profusion.