Cars & Forecasting: Speed or Accuracy?

During 2020 lockdown I discovered a specialist television channel called Motor Trend which presents lots of shows about people mucking about with vehicles. I'm particularly fond of the ones about old vehicle restoration, usually against the clock, with tight budgets, in a limited space, with incompatible parts, inappropriate tools, unreasonable expectations and characters with well, character.

There are loads of car shows you can watch from Rust Bros (on Netflix) to Flipping Bangers (on Prime). The best one though, is RoadKill. It started as a YouTube show in 2012 with David Freiburger and Mike Finnegan throwing darts at a map, then flying to that location (El Paso) to buy a clapped-out car to drive back to LA. The show gathered momentum, an army of fans and has developed over 12 seasons into a much more sophisticated show with multiple offshoots.

Roadkill's central premise is to purchase a piece of automotive junk - something that makes zero financial sense to own let alone fix and then make it go a long way (1000's of miles) and/or make it go very, very fast. There's very little preciousness about the vehicles, they are rarely restored or made what you might call "good". The intent is to go on a journey, see what fixing needs to be done on the way and make it go super quick on a drag strip.

Sometimes the idea is do do a daft thing, like swap an engine on the roadside or in a supermarket carpark with cheap tools, express delivered parts and zip-ties holding everything together. Sometimes it's to just go somewhere and do something that they've always wanted to do.

Can you see where I'm going with this yet?

There are over 100 episodes (and counting) and there are two reasons why I have enjoyed watching all of them. Firstly, it is fun watching the team behave badly with cars. There are lots of stunts, sayings and silliness for the sake of it. Secondly there are simply loads of parallels with demand planning and, if truth be told, Life. Every single episode provides lessons to learn:

• About being roughly right and moving on. Get there and get going again. Moving forward is better than stopping for any great length of time. Sometimes just good enough is good enough. Getting to the end point as directly as possible, even if it means not doing things "right", is occasionally (frequently) the thing to do.
• About making decisions and learning. Even if sometimes, the learning is not necessarily understanding. The very best experience and knowledge can still leave you stumped as to why things don't work as expected. Never mind, revisit the first lesson.
• About being creative with solutions. Think laterally, literally & lovingly. Ask questions. Take things apart and put them back together. Laugh at outrageous options and why not try them? The process of implementing a daft idea will open up new skills and possibilities.
• About how sometimes just good enough is not good. If doing things "right" is more important than anything else, even if that means not reaching the goal in time, then do it right.
• About enjoying the journey, even if it is not enjoyable. A new baseline (good or bad) is a very valuable thing to have. You can push harder or stop sooner when you have bigger baseline and trend line data.

The early episodes of RoadKill were sketchy, spontaneous and fresh. Of course, they were scripted and planned, but the team could do whatever they wanted, and it shows. As the programs progressed, they became tighter and more focused. The production became slicker, quicker and smarter even if uncertainty stalked as every mile was travelled and each zip-tie was added.

Speed!

RoadKill loves a burnout. This is when the rear wheels are rotated at maximum speed while the car is held stationary, and the aim is to create colossal amounts of tyre smoke. Why? It stinks, it looks spectacular, and it became a trademark task in the show. How fast can you go going nowhere?

Making a vehicle go faster uses two approaches: lose weight and add power. Losing weight is easier if you're not precious about the car: chuck out the rear seats, cut the roof off, take out the windows, remove the air conditioning. Adding power comes from tuning the engine or swapping for a more powerful one. Maybe add some chargers (super or turbo) or a squirt of nitrous.

Making a Forecast Cycle faster uses two approaches: lose weight and add power. Losing weight means removing process steps, cut out some of the dataset, throw away the spreadsheets. Adding power comes from training your planners, applying some more resources or swapping the system for a more powerful one. Maybe add some casuals, automation and a squirt of machine learning.

Accuracy!

RoadKill loves a course challenge. This is when they take a few cars for a competition in wholly unsuitable terrain. Maybe a volcanic bed, a frozen lake or a sand dune. Two clapped out cars tuned for racing lines and survival of the fittest.

For a vehicle go more accurately is for it to travel without unwanted deviation. To be able to nail the apex of each corner with balance and finesse. The easier it is to drive, the faster the time will be. Usually. Making a car go more accurately depends on many different factors but stability with flexibility is vital conundrum to resolve. A rigid frame, variable suspension and the right tyres using 4-wheel drive is a good start.

Making a forecast more accurate depends on many different factors but stability with flexibility is a vital conundrum to resolve. A standardised process with variable segmentation and exception management is a good place to start. Maybe add flexible hierarchies & forecast horizons with refined accuracy measures using older historical forecasts.

Because Planning.

Roadkill has a number of saying that occur frequently during the shows and some of the best ones are:

• "Because Roadkill" (usually when something daft has happened or is about to occur)
• "It's Mint!" (description of a car after the first inspection even though the vehicle will be only good for scrap.)
• "Best day at work, ever" (Usually when a catastrophic or bizarre learning event has happened but sometimes when a fabulous achievement or event has occurred)
• "It'll work flawlessly" (usually when a not-good-practice bodge has been applied that will almost certainly fail at some point)
• "Keep lowering your standards until you achieve a goal" (usually after the aforementioned bodge has been bodged again and again)

That'll Buff Out!

Do you want speed or accuracy? If you want both then the ultimate possibility of both will likely suffer. The ability to go faster or more accurately depends on the vehicle you are working with. A drag car needs to go unbelievably fast in a secure straight line. A rally car needs to go unbelievably fast in constantly undulating and changing conditions. Put either of these cars onto the wrong track and they will fail against the competition. Neither is incorrect, they are both tuned for their task.

Do you want forecast speed or accuracy? The ability to go faster or more accurately depends on the data, systems, process and resources that you have as well as the industry, culture and ambitions of your business. If you want both then the ultimate of both will likely suffer. Know your destination. Know your vehicle. Know your capabilities. Apply the right methods. Review & repeat.

Regardless of whether your journey is starting in an old inherited jalopy (spreadsheet & legacy ERP) or is already underway with a new hybrid (on-premise & spreadsheet & cloud) or maybe a brand new electric (cloud) you should be prepared for a bumpy but enjoyable journey.