Imagine you are starting your morning on a Tuesday. You’ve got a cup of coffee and everyone is rolling up to work so you can push out a big order by the end of the week. Just as you’re leaving your office, you hear the phone ring. The call is for a big order that’s way above what you forecast for the week.
What do you do?
You’re not staffed enough to handle it right now. You could pay extra to get your supplier to rush extra material, but you aren’t sure that you could produce it all and ship on time. This example is a common type of tough situation created by unexpected variations from your demand forecast.
No matter how good a forecast is, there will always be surprises and variations in actual demand.
Veryable works with companies who face this type of situation every day. We connect businesses to workers on demand through our labor marketplace, enabling managers to be certain they can adapt to changes like the one in the story above.
Our on-demand labor solution goes a long way in addressing the shortcomings of demand forecasts, but we acknowledge that there are many other ways you can improve your operations besides using Veryable. We’re putting this information together so you can find what works for you.
In this article, you will learn the most common problems with demand forecasting and how to solve them so your operation can stay flexible and competitive.
Common demand forecasting problems
- Forecasting demand too low
- Forecasting demand too high
- Supply chain dependencies causing last minute changes
- Forecasting demand too early
- Relying too rigidly on forecast demand
Solutions to demand forecasting problems
- Forecast demand frequently and for short periods
- Understand which steps of production you can stagger for flexibility
- Have a plan for responding to expected variations
- Have a plan for handling extreme variations and edge cases
Common demand forecasting problems
Forecasting demand too low
Planning to meet demand that’s lower than what actually needs to be met will leave you scrambling to catch up. When something unexpected happens and you don’t have enough materials and labor, you end up shipping orders late because you have to scramble to find workers and the things you need to make the products.
This is bad for the customer and bad for your operational metrics. Scrambling to meet unexpected demand costs money in the form of finding additional workers or paying existing workers overtime. And if you don’t spend to meet the demand, you end up sacrificing your lead time or impacting on-time delivery. If you try to save your on-time delivery by paying extra on shipping, you end up with yet another cost.
Most companies intentionally avoid forecasting demand too low because they cannot find additional materials and workers fast enough without big costs.
Traditionally, temp labor has been used to meet unexpected demand, but there are problems with temp labor you should be aware of if you plan to use it.
To avoid the problems created by forecasting demand too low, you should find ways to keep up with demand in a pinch. This means identifying sources of labor, like on-demand labor, and having protocols in place for quickly ramping up your operations. For example, you could run an extra shift or focus all activity on meeting the urgent order.
Whatever you do, make sure there are plans in place for the inevitable variations that come with forecasting demand.
Forecasting demand too high
Planning to meet demand that is higher than what actually materializes will result in overstaffing and excess inventory. This leads to overtime costs and potentially storage costs for the extra materials you have on hand.
Many companies choose to prepare for the higher side of their demand forecasts and accept these costs, because the alternative is scrambling to catch up as mentioned above.
But forecasting demand too high is just as costly because overtime costs add up quickly and worker morale dips when they have to stand around with nothing to do. Lower morale leads to turnover, and that puts your company back into scramble mode to find workers as the cycle of costly dependence on inaccurate forecasting continues. You also run the risk of overproduction, which ties up money in inventorying finished goods.
To avoid the problems created by forecasting demand too high, you should know what to do when there are too many workers and materials available. Plan for how you can keep workers doing value-add work if they’ve already been scheduled, or identify workers who are willing to drop shifts when there just isn’t work to be done.
You should also find a way to handle any excess materials, like cost-effective storage or a way to quickly repurpose them for other projects.
Supply chain dependencies causing last-minute changes
The companies on either side of you in the supply chain impact your ability to keep promises. This is amplified by any inadequacies in your demand forecast.
Imagine that your supplier ships you something late, and you had forecast demand at 100 units for the week. You only have enough materials on hand to produce 80 units, so 20 units are on hold until the supplies arrive.
Now you can ship those 20 units late or you can scramble as soon as the order arrives to produce them and ship on time. In this situation, if you take the order and want to ship on time, you are forced to find extra workers or work your current employees around the clock once the supplies arrive.
This is the kind of rock and a hard place decision that the complexities of a supply chain can put you in. Sound familiar?
Demand forecasting only adds to the frustration when it is inaccurate. Imagine forecasting 100 units, then your supplier ships materials late, and on top of that you get an order for 10 more units. It’s like gas on a fire.
To avoid getting burned by the complexities of the supply chain, you should find ways to give yourself more flexibility in responding to demand. To do this, you’ll need to tighten up your demand forecasting to a shorter time period and conduct it more frequently. That way you can get it closer to the timeline for scheduling your workers, and you can more easily avoid being overstaffed or understaffed.
The closer you are to your actual demand with your forecast, the closer you can plan the work to the actual demand, and the less affected by unexpected changes you will be.
Ideally, you could respond to demand just-in-time. This would require a scalable solution for finding workers quickly, and planning to meet the lower range of your forecast demand. That way you could bring in workers on an as-needed basis when demand changes suddenly. This dampens the effect of supply chain dependencies on your operation.
Forecasting demand too early
Forecasting your demand too early means you’ll almost certainly miss the mark. The further out you plan, the harder it is to be precise with your predictions.
Even if you can start forecasting on a weekly level or more frequently, the mismatch between the timeline workers need and the timeline for forecasting demand creates problems. That’s because schedules for workers are typically decided a week in advance. This means you’re having to guess demand one week out, even if there’s no way to accurately do so.
For some businesses, this isn’t that much of a problem because their demand is very predictable. But for most businesses, the goal is to reach a real-time response to demand. Getting as close to this as possible with forecasting demand is the next best thing.
To avoid forecasting demand too early, aim for as short of a time period in your forecast as possible. For example, forecasting for the week is better than forecasting for the month. Forecasting for the next couple of days is even better if you can get past your labor constraints to do so. The closer to the actual production of orders you can forecast demand, the closer you are to the ideal of real-time response to demand.
Relying too rigidly on forecast demand
Having a plan is great, but sticking to it so rigidly that you can’t respond to the reality around you is a surefire way to get your business into trouble. Rigidity costs you time, money, and opportunities.
No matter how you forecast demand, it will have some degree of inaccuracy. It is common to predict demand and claim a 95% accuracy rate, which means there is an acceptable range of variation of 5% from the forecast level of demand. This can translate into big costs related to overstaffing or being unprepared even in the best case forecasting scenario. It is more likely that the forecast isn’t even 95% accurate.
Depending too much on forecast demand can make it so that companies won’t look at the situation unfolding around them and make a change to respond to the situation.
Too often companies stick to their forecast even when it doesn’t make sense to continue on that path. They will keep too many workers and pay overtime because their forecast was too high, or they’ll stay in scramble mode for weeks while understaffed because that’s what the forecast said they’d need to do.
Sometimes, it is impossible to forecast demand. In some cases, you could be given an order on Sunday night with a must-arrive-by date of Thursday. It is common in these agreements for you to be fined some percentage of the sales value of the products by the customer if you don’t meet this date, so it could cost your company real money to depend on the ability to forecast instead of responding in real time to demand.
So what are the options for avoiding being overly reliant on forecast demand? How can you add flexibility to your company and respond to demand quickly, no matter what direction it goes?
Solutions to common demand forecasting problems
Forecast demand frequently and for short periods
Forecasting demand over short time periods and conducting forecasting more frequently reduces the chance that you overextend your predictions, causing you to overestimate and overspend or be unprepared for demand.
In Lean methodology, a core principle is establishing flow, meaning nothing is created until a customer orders it. Establishing flow is meant to reduce wastes like excess inventory and delays by responding just-in-time to demands. When you forecast your demand for shorter time periods, you’re closer to the ideal of just-in-time manufacturing. This eliminates the waste of paying overtime when you’re understaffed, or in the case of being overstaffed, having to pay people to do nothing or busywork.
Understand which steps of production you can stagger for flexibility
If you have a longer lead time, sometimes you can speed up your response to demand by understanding which parts of your process you can do before an order is placed.
For example, if your product is customizable in some way, you could produce the item until it is just before the point of customization. This way you have a pile of items ready to be customized when orders come in, making it easier to respond to variations in demand.
Have a plan for responding to expected variations
No matter how great your model used in forecasting is, there will always be some level of inaccuracy and uncertainty in the forecast. That’s why you should always have a plan for the expected amount of variation from the forecast.
That means you should plan for what happens if too few orders come in as well as what to do if too many orders come in.
One solution to receiving less than the expected orders would be to plan to the lower bound of your forecast. That way you are closer to the headcount and amount of materials you actually need.
Create a list of tasks that you save for slow periods so that workers will always have something to do that adds value to the production process. Maybe there is even another production project that you could squeeze into the extra time you have.
On the other hand, you also need a plan for when more than the expected orders come in. To put a positive spin on too many orders coming in, you could plan to only have enough people working to meet your minimum forecast demand and invite workers as the extra demand rolls in. Solutions like on-demand labor allow you to find workers within a day, so you could quickly respond to extra orders by adding more labor to the equation.
Have a plan for handling extreme variations and edge cases
Make sure your solutions for handling variations are scalable. Maybe you’ve planned for what to do if demand comes in at 5% more than you projected, but what would you do if it came in at 15% more than you projected?
It’s becoming increasingly obvious that in the Internet Age, demand can become volatile overnight for some businesses. Be prepared for extremes so you can have peace of mind, no matter how volatile demand might be.
You should also have a plan for what you’ll do if more than one thing goes wrong at once. For example, what would you do if an important machine breaks, a star employee takes vacation, and a big order comes in all at once? It’s unlikely, but it couldn’t hurt to be prepared for this type of situation.
Avoiding over-reliance on demand forecasts with on-demand labor
You can avoid much of the frustration that comes with relying on demand forecasts by adopting a more agile approach to planning. This involves having a reliable way to find workers on short notice, so you can keep costs low with confidence that you can respond in real time to demand.
New ways of planning, when combined with on-demand labor can empower you to rely less on guesswork and respond quickly to demand. This article on capacity planning for manufacturing is the next step in your quest to solving demand forecasting problems.