These are well-known examples of forecasting failures, but there are loads more. There’s no escaping that businesses are generally bad at the whole thing. Nearly 80% of sales organisations miss their forecasts by at least 10%. And perhaps just 1% of companies hit their forecasts exactly. You have to wonder what’s going wrong.
Why should you forecast?
Judging by all the failures, it is a fair question to ask: why bother forecasting? Well, in truth, effective forecasting offers so many advantages that a company would be foolish to ditch the idea of making forecasts. Moreover, thanks to AI, forecasting is generally getting better, with the technology reducing errors by 20-50%, according to McKinsey.
Here’s a short list of some of the benefits that business forecasting offers:
- Improved decision-making –
Companies make better-informed decisions with data-driven insights about future trends
- Greater operational efficiency –
Forecasting enables businesses to allocate resources more effectively, predicting demand to optimise inventory levels and production.
- Enhanced customer satisfaction –
It predicts customer demand and behaviour so products and services are tailored more accurately to the market.
- Risk mitigation –
Using forecasting as an early warning system, companies can anticipate potential challenges and disruptions.
What can be forecasted? – 5 areas of focus
While, in theory, there’s no boundary on what a business can forecast, life, as we all know, comes with constraints, and companies need to prioritise. Here are five forecasts covering the basics that will help a business operate more efficiently:
- Sales figures (the obvious one) –
These forecasts impact business operations from many different angles. With a sense of what sales will be in the future, it’s easier to plan resource allocation, cash flow, inventory, marketing, etc.
- Churn –
Forecasting churn rates can help a company remain on the front foot when retaining existing customers. Acquiring new customers can be up to 25 times more expensive than retaining current ones so that forecasting can play an important role in optimising business spend.
- Stock –
Every business wants to avoid unnecessary stock costs, making forecasting essential. Forecasting helps minimise overstocking and stockouts and avoids the expense of rush orders.
- Traffic (onsite and instore) –
Being able to anticipate traffic flows helps companies prepare for peak periods. That could be adding IT capacity for spikes in website traffic or increasing in-store staffing levels. It can support a better omnichannel experience for customers.
- Turnover rates –
Knowing how many employees will leave within a defined period, a business can calculate the likelihood of a staff shortfall or surplus and start addressing issues before they become actual problems
Common pitfalls in forecasting
At Ipsos Jarmany, we’ve witnessed plenty of forecasting pitfalls during customer engagements. The good news is that the same pitfalls occur over and over again, which means the same mistakes are being made, which means the problems are easier to identify.
Here’s our list of the most frequent issues that we come up against and how to avoid them:
Know what you’re trying to forecast
Identify the purpose of your forecast, i.e., sales figures, turnover rates, etc., and determine if you have the relevant data. Get focused on a particular variable, such as historical sales data, to simplify your forecasting process. The reality of focusing on one forecast at a time leads to more accurate and easy-to-interpret results.
Feed in good historical data (quality, quantity, etc.)
Be aware of unexpected results in your data or one-time spikes, which could be due to legacy business processes, historic bugs, or sudden, massive increases in sales. Ensure your historical data is accurate and clean: check for errors, duplicates, and any outliers that might skew your results. Collect enough historical data to capture a wide range of fluctuations and incorporate data from multiple sources. These could be internal data, industry reports, and customer surveys.
Leverage human historical insight (when possible)
Don’t put all your faith in the tech stack. Sure enough, forecasting nowadays is tech-driven, but that doesn’t rule out the part to play by us humans. It’s always good to surface any historical insights from employees who might be able to explain, for example, why there was that spike in sales six months ago, which may be hard to explain to the forecast model.
Set yourself a time limit
Ten-day weather forecasts are much less reliable than three-day forecasts, yet today’s meteorologists have access to a whole bag of sophisticated tools and supercomputers. Bear that in mind when planning your forecasts, and don’t look too far into the future. Three years is probably about enough.
Working as a team
No matter how long the forecasting team works on its forecasting models, it will never achieve great results unless there’s buy-in from the rest of the business. Forecasting must be a collaborative process where assumptions are tested and validated in review meetings. The more diverse the perspectives and qualitative insights made available, the more accurate the forecast.
How should you view forecasting – assistant or seer?
As we’ve seen with forecasting, it can go wrong, and mistakes are often made. But you can guarantee that Coca-Cola will always forecast no matter the number of wrong calls it makes.
And anyway, the development of predictive analytics, now supercharged with AI, have upped forecasting’s game substantially.
Businesses like Costco, a major retailer in the US, is heavily invested in predictive analytics to forecast future demand for different products. Plus, Nike has acquired analytics companies over the years to help it forecast consumer behaviour in order to improve customer acquisition and enhance accuracy of its sales.
But leveraging forecasting to the max isn’t merely a question of getting the IT right. While forecasting is an amazing assistant, it’s not sound kind of personal seer . Human experience and judgement will always be an important part of the forecasting equation, and what the technology tells us shouldn’t be the be-all and end-all.
Remember, forecasts provide a framework to support planning and decision-making, they don’t make decisions. They are based on historical and present data and assumptions that might not consider sudden market changes. If we treat forecasts as flexible tools, we can develop strategies to continually improve them but keep an agile mindset in case you need to pivot.
The ideal forecasting team
You want team members to have business acumen and communication skills. You want them to have financial knowledge or expertise in the industry being forecasted. Plus, there’ll need to be good communicators to work effectively with the broader company and ensure its buy-in.
Proficiency in using tools like Python or R is a must, as is a grounding in Excel or PowerBI, which has native forecasting functionality and visualisation tools built in. You’ll also want team members to understand statistical techniques, such as regression analysis and time series analysis.
Finding the right talents and building your forecasting team can be done, but it takes time and resources. At Ipsos Jarmany, we’re helping customers fast-track their forecasting capabilities by providing them with the needed experience and expertise. Our team helped a global consumer technology company transform its forecasting operation. The company reduced its client’s sales forecasting process from 2 weeks to a matter of hours while gaining more granular insights. We also helped increase the sales team’s efficiency, reduced the risk of human errors, and provided a solution to adapt dynamically to an ever-changing market.
Get in touch
Our forecasting experts use their business acumen, communication skills, and statistical skillsets to give customers the forecasting capabilities they require to thrive in today’s world. In each instance, our engagements started with an open discussion about how clients could integrate forecasting into their business toolkit, and our relationships have expanded from there.
If you’d like to learn more about how we can help you develop a successful forecasting capability, get in touch with us today.
