Today, when Business Intelligence is a standard at most large companies, decision makers have started to rely on important business metrics in their everyday operational work. The internally collected, processed and provisioned data is not only used for building informative dashboards but is also the input for more advanced analysis applications like demand predictions, capacity planning, or route optimizations. In thea world that undergoes little short term change, this works very well for many cases even without explicitly including data on external factors. The ever-repeating patterns of seasonality that are caused mainly by the cycles of daytime, weekdays, seasons or regular events are well captured in the historical data that is generated by the normal business operations. Macro trends that cause severe shifts are harder to capture, but these are usually rather slow and therefore can be accounted for with more old fashioned formats like traditional market research or quarterly strategy meetings.
The recent pandemic, however, has mixed up our whole world quite significantly. Many stores and even whole cities were closed one week to the other which led large parts of demand to shift from offline to online within a few days. Also our demand itself changed for various reasons. As most events are cancelled, we have less reasons to buy and wear chic apparel and instead rush for the remaining sources of fun and distraction, leading to soaring sales in video video games, (e-)bikes and apparently toilet paper.
These short term shifts immediately raise the question of long-term effects. Will we continue to buy online, pay digitally and cook at home? Or will everything go back to pre-crisis levels when the virus is somewhat contained? No matter what the answers are, one effect is already clear: just using our historical data to predict the future won’t work anymore, as the past is most likely not a good blueprint for the future anymore. Like our favorite vinyl record that got a deep scratch across its whole surface, our precious time series are damaged heavily, and therefore, getting valuable insights from the data will be even harder than it was before.
But the good news is that we don’t just have to accept our destiny and go into the future blindly. In fact, making good predictions is still possible in many cases, we just need to get a bit more advanced with our analytics. To be precise, we need to give our models more information about what is happening in this constantly changing world around us. You can compare it to driving a car with closed eyes. It might work as long as the street is going straight ahead and weather conditions are good. But as soon as the street gets curvy and the conditions get rough, you need to open your eyes. And it is not enough to look on the speedometer! We need to see what is happening around us, if there are potholes or a big truck heading into our direction.
That this strategy works was proven very well by Procter & Gamble who started building their fully dynamic and data-driven supply chain already many years ago as a reaction to the devastating impact of natural disasters. They digitized all relevant processes, connecting the various ERPs of their plants, distribution centers, and other facilities, and implemented a system for data-driven (semi-)automatic decision making. The input to this system is all the internally generated data including plant capacities, stock levels and product information. But in order to work well, the system also uses a wide range of data sources covering external factors like weather, traffic, or market demand.
By constantly optimizing their internal system to optimally adapt to the outside world, P&G was able to outperform their competition significantly throughout the Corona pandemic. While one might think that for a company producing toilet paper and cleaning supplies it is a given to perform well in a pandemic, the direct competitors obviously had a harder time adapting to the fast changing environment and thus were only able to hold their pre-crisis levels. P&G, however, was able to increase their 2020 sales by 6 % and their profit by 13 % which led the stock price to rise by more than 17 % (November 2020).
Given this very successful real-world example of a data-driven company, the natural question is how we can get there as well. In the end it doesn’t matter if we are talking about a human decision maker or an “AI”. In order to make well-informed business decisions, we need sufficient information about what is happening around us. Traditionally, we as humans get this information from traditional sources like the news, books, courses, market reports or industry experts. Machines need the same information in order to perform well - with the difference that it needs to be machine-readable, i.e. represented in a numerical way. Independent of the format, this information can be separated into four categories.
Data on our target markets, including customer information, market trends, prices, quantified demand predictions, external effects (e.g. weather, events, or Corona). This information helps to estimate what the market will demand in the future.
Data on our supply markets - the markets we source the materials, products and services required for our business operations from. Relevant information includes data on existing and potential suppliers, new technologies, prices, or delivery routes.
Data on our competitors like financials, performance indicators, existing and new product lines, market shares, strategic decisions, geographical focus, suppliers and customers, or decision makers.
Data on external factors that could negatively impact our business operations, like natural disasters, reputational risks, business partner bankruptcy, fraud, or regulatory changes.
By gathering relevant external data and combining it with your internal data, you can build an accurate and constantly updating data foundation that enables you to make better strategic decisions and develop more robust forecasting models that perform well even in times of constant change and uncertainty. While it is not an easy goal to achieve, this is the first step to building a cognitive and fully adaptable supply chain that allows you to react faster than your competition and thrive when others try to keep afloat.
Luckily, you don’t have to solve your data challenges on your own. The number of free data sources, data vendors, and data-driven solution providers is constantly growing so your options are manifold. No matter if you need a turn-key solution or want to build your own, powerful data foundation, we support you as your end-to-end partner for external data. Just get in touch with us via firstname.lastname@example.org or use our contact form.