What can the 2016 US Election teach us about forecasting?
(This was first published on Linkedin here)
The world changes everyday. Normally the changes are small enough that you don't notice them. Sometimes a change happens that's a little more substantial - one of those happened yesterday in global politics.
It became very apparent the result caught a lot of very intelligent, well educated, experienced "experts" by surprise. Why is that? Why can't the "experts" forecast the outcome? It's a deeply unsettling feeling.
This morning I've seen several articles of people attempting to predict what may happen given the seismic shift. Einstein's definition of insanity is useful here - "doing the same thing over and over and expecting a different result". I'm going to buck the trend here - there are no predictions here on what the election of Donald Trump as US President means for the world.
Clearly the rule book has been thrown out and we are in unchartered territory. So was Christopher Columbus and he did ok. So how can good forecasting help you in an uncertain business environment?
Forecasting is built on the fact there are "cause and effect" patterns. If you can understand those patterns from the past, and more specifically, what were the drivers (the cause) of those patterns, you can model what the future (outcome) may look like. You can also use those models to understand how a small change in one of those causes could change the outcome.
A good example is weather forecasting. How does the weather bureau (and they do get it right more often than you may think) know what the weather's likely to be tomorrow? They look at a bunch of factors including wind direction, humidity, high pressure systems, low pressure systems, and how they interact. In Australia, most weather patterns move from the west coast to the east coast which also helps give the east coast some warning as well.
Once you get more than 2 to 3 days out, the reliability level drops substantially. Why is that? It only takes a very small change in one of the drivers (cause) to substantially change the outcome (effect).
Another set of forecasts are those done by Australian Treasury for revenue projections so that long term spend planning can be done by the govt. Those forecasts have not had a good track record in accuracy for quite some time now.
Anyone who's ever done any modelling understands the frustration of seeing a news journalist use the fact that events didn't transpire the way they were forecast as a way of ripping into someone & suggest they're dishonest. It's one of the many reasons I'd never go into politics, that's not my idea of a good time.
It's also been used as a way of trying to discredit climate scientists - the global warming models are inherently uncertain due to the number of causes and the variability in how they interact with each other. The communication methods that make sense to the scientific community do not translate well into the mainstream news cycle.
So why forecast? If it doesn't give you certainty on what's going to happen, why bother?
Forecasting in business is a very effective tool to do three things:
a. Direction - which way should I go?
b. Focus - where should my investment dollars go to grow my business?
c. When do I need to change course - what are those drivers which, if they change, mean you need to respond and change direction quickly.
Think of the rudder on a large cruiseliner. It manages to keep a very large vessel on course with substantial pressures coming from the ocean to push it off course. On board is a navigator to plot the course, to watch ahead and figure out when to change course. Something may change very rapidly and they need to respond and change directions quickly.
Direction - which way should I go?
Do I stick with my current customer segment, or expand into another area. Do I add additional product or service offerings, and if so, into what markets.
What would my revenue and margin (revenue less cost of delivery) look like under each of these scenarios.
As and when we don't meet the forecast, why not? Is it due to number of new customers, product mix, retention of customers, pricing? These comparisons allow you to figure out where you need to tweak as you go.
This works well in a "business as usual" market. When that's not the case, focus and adaptation become even more important.
You have limited time & money. Where you do you spend it? You can only do so many initiatives at once that will grow your business.
Focus is about picking those areas which will have the biggest impact.
We went through this for the budgeting process for one of my clients earlier this year. We picked three main areas of investment, one to substantially grow the business and two to "future proof" the business.
For various reasons revenue was not hitting the forecast we'd done. We re-did the projections and protected two of the three investments, due to their importance to the long term health and vitality of the business. We could do this quickly as we were focused on our priorities, and also understood what the drivers were.
Adaptation - when do I need to change course?
What is it that can change in your business model that would substantially change the outcome?
Is it the length of time customers stay with you? Is it the amount they spend? Is it the cost of a key component going up? Is it the loss of a key staff member.
Often these are the things which can take you by surprise as they're based on an assumption which held true for a period of time but is no longer the case or was always fundamentally flawed. What's happened over the last week is testament to a flawed set of assumptions used by the "experts" on how people behave and how this translates into voting patterns.
What is it within your business that, if it changed, would turn your business model upside down?
Can you scan the horizon ahead (like the navigator) and, if it happens, change direction fast enough?
Are you making assumptions in your business model and assuming they're facts? For IBM it was that "no one was ever fired for buying an IBM". With cloud computing, that's not a buying decision that's coming up a lot anymore so is no longer a valid assumption in developing marketing messaging. IBM has adapted into cloud computing, cyber security and other areas. Their revenue is dropping, can they adapt quickly enough? Can they turn the ship in time?
Forecasting allows you to see the road ahead and change course when you need to. As the driver, only you can make the decisions to change when needed, good forecasting is part of your toolkit so you're not doing it blind.