We humans are amazing.
It has taken about seven decades of cooperative technological endeavour to bring robots even close to the skill of a human being walking on rough ground — something that tiny babies can master in a month or two. We can achieve this because our senses provide continuous feedback from every movement and adjustment. If that baby had to go through all the calculations made by robotic developers in learning to emulate walking … Well, just forget it.
Most of us learn these, and other motor skills, well enough to get through the competitive pressures of everyday life, and that is sufficient. But some people go on to amazing heights — just think of the split-second precision, timing and judgment required by trapeze artists or tennis champions. Our ability to gain insight and foresight from continuous feedback is unlimited and can, with training, lead to unbelievable levels of performance.
What would happen to a business if it could be guided by continuous forecasting instead of the traditional year-end forecasts? The latest enterprise performance management solutions are getting closer to that ideal and — as business learns to use it — the benefits are looking impressive.
The Need for Agile Decision-Making
Manufacturing companies that need to budget for next year’s raw materials plan their purchases and targets well in advance in a relatively stable trading environment. So performance management for a typical CFO has long relied on a fixed system of annual plans, targets and forecasts. But manufacturing no longer dominates Western business, and we face a far less predictable business environment where startup rivals can appear from nowhere with new business models, like Uber and Airbnb. Instead of steady evolution, today’s business faces unpredictable, discontinuous changes, and customers can rapidly switch allegiance as well as demand new levels of service.
Making wise decisions fast enough is no longer possible using yearly, or even monthly, reporting. It can take a hundred days and 20 percent of management time to complete an annual budget; meanwhile, operating and market conditions deviate from the budget assumptions until performance reporting becomes meaningless and opportunities get missed.
As a reaction to this dynamic business environment, annual planning cycles are giving way to more frequent business reviews, supported by rolling forecasts that allow people to spot trends, patterns and possible disruptions before their competitors do. Today’s managers need to work with continuous forecasting and planning cycles based on key performance metrics in order to stay agile and competitive.
Continuous Forecasting Technology
New tools are being developed to deliver the necessary near real-time reporting and forecasting — and not just for the business community. As an example, consider how the rise in urban traffic makes it increasingly difficult to forecast how long any journey will take. So, for an important appointment, it becomes necessary to allow a lot of extra time. You may have been forewarned to avoid one main road because of road works, only to find that everyone else has been told the same and now the side roads are also jammed. As a solution, we now have Waze — the community-based traffic and navigation app that is continually updated by its own users to warn others of bad traffic, accidents and hold-ups. With an app like Waze, drivers get insight into traffic conditions on their route as they actually happen.
Apps for weight watchers provide another example of the transition from hindsight to real-time data. The old way was for a person to weigh himself and get a measure of how well he had been managing his weight in the previous few days — so he can try to do better next time. But the latest iPhone apps invite an individual to take a picture of the plate of food served to him and get an estimate of its calories, while fitness trackers will tell him how many calories he is currently burning — again, this is real-time data that helps the individual make informed decisions on the fly. The latest techniques for continuous monitoring of blood sugar levels provide similar benefits for diabetic patients. Even with the latest portable self-measurement devices, the user only gets a record of the level at the time of measurement, while continuous measurement systems give real-time measurements with algorithms to determine trends, recommend action or sound alerts to the user and parents, doctors or other parties as needed.
Similarly in business: Today’s more agile organizations are going beyond their traditional enterprise resource planning (ERP) solutions to manage planning as a continuous, event-driven process that is not constrained by the financial year. It may still be appropriate to set regular strategic reviews or to hold a planning meeting in response to some major change, as one Fortune 50 company CFO explains: “Once the year is under way, we review performance twice a quarter. We ask questions about how we are doing, what’s changing in the marketplace, what are the new opportunities that have arisen, and so forth. We might then produce a new estimate based on the latest knowledge. Plans never work out the way you expect, so you have to adjust as you go.”
Adjusting “as you go” not only requires near real-time data on key performance indicators, it also means making that data much more widely available across the organization. Let’s say there’s a special offer posted on a company website that could trigger a surge in orders. Traditionally, the website manager should inform the product manager and the product manager should make enquiries to ensure adequate stocks and so on up and down the hierarchy and chains of command. Unless this is done absolutely consistently according to best practice, there is always a chance that someone along the supply chain gets overlooked — maybe the goods are in stock but a forklift driver is off sick. Instead, an automatic alert can be broadcast to all relevant parties, warning of a likely surge in demand and reducing the risk of human error.
That last example touches two key points: 1) an alert, and 2) only sent to relevant parties. Today’s automated systems are not programmed to swamp every busy staff member with a flood of irrelevant data. Someone in charge of stocks can be forewarned about special offers and promotions that could cause a surge in demand, but need not know about every other website update. So the best solutions allow each user to tailor her own dashboard to deliver the type and format of data she really wants.
The Rise of the Algorithm
Forewarning of a surge in demand is useful, but how big a surge can we expect? Here, the most advanced reporting systems provide for fixed or self-learning adaptive algorithms. So a special offer being launched on a website might register a number of factors, such as the magnitude of the reduction, the size and distribution of the banner ad, the time of the announcement, current purchasing rates, competitors’ activity and so on. Then an appropriate algorithm could make an order of magnitude prediction of the expected sales increase based on these figures and on past experience.
The use of such algorithms could be the “next big thing,” according to Peter Sondergaard, senior vice president of Gartner Research, speaking at the 2015 Gartner Symposium: “Data is inherently dumb. It doesn’t do anything unless you know how to use it and act with it. Algorithm is where the real value lies. Algorithms define action.” In particular, users can generate algorithms based on key performance indicators (KPIs) that can include both internal data — such as the number of customers currently on the company website or in its retail outlets — and external data such as the weather, competitors’ activity, inflation and other economic data. In doing so, the algorithms can actually reduce the data flood by selecting only those KPIs and ignoring lesser factors unless they reach a critical level. Adaptive organizations respond quickly because they respond to important, relevant data.
Many more people across the organization have something useful to input into such a system, and many more can make efficient decisions given appropriate and relevant access to the resulting performance data. Automation allows the business to focus more on performance and its drivers, instead of just gathering data in the light of assumptions made in the last budget.
According to the CFO of a Santa Barbara phone-call tracking business, it means “you can capture data that’s not always necessarily financial in nature and bring it into reporting or analysis tools, and that’s something that is really, really new. Companies are seeing that there’s huge benefit to that in terms of predictive analysis. … It forces the finance group to evolve and become really much more focused on analytics.”
Another reason to replace annual budgets with continuous forecasting is budgeting demotivates people — no one looks forward to the budgeting operation. It is seen as a bureaucratic procedure that takes time and energy away from the day’s work, whereas continuous forecasting becomes an integral part of normal operation.
When Will This Happen?
Surveys suggest that most large organizations are serious about managing strategy rather than numbers, and want a suitable planning and forecasting model that offers immediate visibility without drowning them in targets and measures. Continuous forecasting is still new and we are on the learning curve. But go-ahead companies are already reducing the flow of manually entered spreadsheets, and replacing them with automated reports that ensure consistent data across the organization. One such — the leading global facilities management company ISS — was able to weed out thousands of redundant, error-prone spreadsheets and accelerate its budgeting cycle in the first year, claiming it was “creating time instead of chaos” with benefits seen far and wide across the organization.
Sports equipment company Mizuno USA is moving from annual budgeting to a more continuous forecasting process that cuts days off its financial month-end close reporting. Its senior financial business analyst, Seth Higdon, works with the supply, marketing team, sales and product teams providing reports generated by a real-time analytics and reporting solution. The immediate result has been improved collaboration and, he says, “better real-time decisions from the guys getting these reports … This has really made Mizuno USA a much better-performing company.”
For the Womble Company, based in Houston, Texas, the emphasis has been on improving response to customer queries with real-time reporting. And another Hubble customer, Zuffa — the company managing the UFC (Ultimate Fighting Competition) — is using real-time reporting to be more responsive to business opportunities and increase profitability.
In these, and other examples, the shift from historical data and manual spreadsheets — toward automated access to real-time data plus customized alerts and analysis of trends — is already under way and showing positive benefits. But there is also a general feeling that this is the way forward and there is much more that can be achieved in terms of continuous forecasting.
The focus so far is on agility and competitive advantage, but there are also important implications for accountability and compliance. While traditional spreadsheet data passes up the business hierarchy, it is subject to manual revision. This can lead to problems, such as a department failing to meet targets that had been revised unrealistically elsewhere in the organization. Automated reporting not only passes out consistent data to everyone, it also keeps a full record of that data and any revisions that would provide a clear audit trail if needed.
Going back to the analogy between continuous forecasting and human learning: Business today is accelerating from a crawl to a walk to a fast jog — but there is still a lot of learning before we can compete at the level of an Olympic gold medallist.