Methods for Supply Chain Forecasting: Preventing Storms, Predicting Trends
Methods for Supply Chain Forecasting: Preventing Storms, Predicting Trends
Weather forecasting and supply chain forecasting have many things in common.
They both make predictions based upon past and current information. Both use hard data and sometimes intuition to determine the outcome. In both cases, you can feel caught out and unprepared if something doesn't show up on the radar.
It is crucial to understand how to forecast your supply chain requirements in order to ensure your ecommerce store's success. It can help you build better relationships with suppliers, increase customer satisfaction, and provide more capital for your business to grow and scale.
We talked to supply chain management, fulfillment and shipping experts to learn how supply chain forecasting can affect your store's next quarter, as well as the best ways to do it
What is forecasting for supply chain management?
No matter if you order whole products or raw materials that must be assembled, forecasting is part of supply chain management. It involves determining when suppliers have the product available and when it should be ordered.
You need inventory to deliver orders quickly and cheaply. "Keeping track of inventory velocity over time requires being able monitor best-sellers, stay ahead of production, even as demand changes," Kristina Lopienski (director of content marketing at ShipBob), a global logistics platform that fulfills direct-to-consumer ecommerce orders.
Some key factors could include:
- Product turnover rate
- For each product or supplier, lead times are required
- Transit times for freight
- Time to receive in the warehouse
- Storage costs
Supply chain forecasting, as its name suggests, is based largely upon analyzing supply. Demand also plays a role. Factors like seasons, trends and global events can all cause spikes or slow sales that can impact inventory control.
Why is supply chain forecasting so important?
To understand that timing is everything, you don't need to be a regular reader of Journal of Supply Chain Management.
Leandrew Robinson is the general manager of mesh logistics at Auctane, which includes ShipStation and ShippingEasy as well as ShipWorks and ShipEngine.
"If supply chain forecasting doesn't accurately down to a few weeks, it can lead to costly ripple effects."
In an age where 67% U.S. consumers expect next-day, same-day, or two-day delivery, products that arrive late at your warehouse or shipping centre won't reach customers on time. This can not only damage your brand's reputation but also cause sales losses. Your customers will look elsewhere if you don't stock it or if it's backordered.
However, inventory that arrives before you are ready can cause warehouse costs to rise or loss if the product has a short shelf-life. This also reduces capital that could be used to expand or improve other aspects of your business.
You may also be charged for incorrect quantities or products. There may be dead stock.
"Stale inventory is sitting in a warehouse collecting dust and accruing fees. You can save such situations by selling at-cost, at steep discounts, or selling in bulk at clearance houses," Nicholas Daniel-Richards, cofounder of ShipHero, which offers warehouse management software as well as shipping solutions.
Supply chain supply chains use quantitative forecasting methods
Quantitative projective forecasting uses historical data to predict future sales. These complex mathematical formulae are based on the assumption that future sales will reflect the past. They are usually performed by computer software.
- Moving average forecast
- Exponential smoothing
- Auto-regressive, integrated moving average
- Multi-aggregation prediction algorithm
Forecasting using moving averages
- ProsSimple
- ConsIt doesn't allow for trends or seasonality
- Ideal for:Items with low volumes
This is one of the most basic methods of forecasting. It examines data points and creates an average series of subsets using complete data.
Moving average forecasting is based on historical averages. It doesn't consider that recent data might be a better indicator for the future. Therefore, it should be given more weight. It doesn't account for trends or seasonality. This method is ideal for inventory control of low-volume items.
Exponential smoothing
- ProsSimple; Takes historical and current data into consideration
- ConsForecasts can be delayed due to the possibility of lag
- Ideal for:Forecasts for the short-term or other non-seasonal items
This method picks up from the place average forecasting stops, but it gives more weight to recent observations. This is similar to adaptive forecasting which considers seasonality.
There are many variations on exponential smoothing, including Holt’s Forecasting Model (sometimes called Trend Adjusted Exponential Smoothing (or double exponential smoothing) or Holt-Winters Method (“triple exponential smoothing”) which takes into account both seasonality and trends.
Auto-regressive integrated Moving Average (ARIMA).
- ProsVery precise
- ConsTime-consuming and expensive
- Ideal for:Timeframes up to 18 months
Box-Jenkins is one method that falls within the ARIMA category. This time series forecasting method can be costly and time-consuming. However, it is the most accurate. It's best for forecasting within 18 months.
Multiple Aggregation Predictional Algorithm (MAPA).
- ProsAvoids over- and under-estimating
- ConsIt is still relatively new and not yet proven.
- Ideal for:Seasonal products
MAPA is a relatively new technique that is specifically designed for seasonality. It smooths out trends to help avoid over- or under-estimating demand. Research has shown MAPA performs better .
Methods for qualitative supply chain forecasting
It can be difficult to forecast supply chain trends for new products or businesses when there is limited data or difficulty obtaining it. You also have the possibility of historical data becoming outdated or less accurate due to global pandemics. This is where qualitative forecasting steps in.
These methods include:
Historical analogies
- ProsIn the medium- to long-term, it may be more accurate
- ConsIn the short-term, poor accuracy
- Ideal for:Similar products
Historical analogy forecasting uses historical sales data to predict future sales. It assumes that a new product will have sales history similar to the one you are selling or a competitor's product. It is less accurate in the short-term but may be more accurate in medium- and long-term.
Composition of the sales force
- ProsIt is very easy to collect
- ConsFair to poor accuracy
- Ideal for:If quantitative methods aren’t possible
This method is sometimes called "collective opinions". It relies on the individual insights and opinions of experienced managers, staff and other members of the team. Harvard Business Review states that panels of this type have a low to moderate level of accuracy.
Market research
- ProsThis provides insight into your target audience
- ConsIt can be time- and/or financial intensive
Focus groups, polling or surveying your target audience are all possible methods of conducting this research.
The Delphi method
- ProsUnbiased
- ConsUncertainty in reliability
This technique involves sending individual questionnaires to experts. After each round, the responses are compiled and shared with other members of the group until there is a consensus. The panel does not collaborate so bias is eliminated.
This method is widely used for long-term forecasting.
Which is the most effective method for supply chain forecasting
No matter what supply chain forecasting method you use, there are bound to be errors caused by assumptions. It's impossible for anyone to predict the future with 100% accuracy. However, you will find that short-term forecasts are generally more accurate than long-term ones.
Our experts did agree on one thing: Qualitative methods are based on opinions of consumers or market experts. These opinions are subjective and less accurate.
Daniel-Richards says that quantitative and trend forecasting, based on hard data analysis, is the strongest supply chain forecasting method. Jokingly, he adds: "The weakest method of supply chain forecasting is quantitative and trend forecasting based upon hard data and analysis."
Why is supply chain forecasting so difficult?
Regulations changing
COVID-19 has caused havoc in supply chain forecasting systems in more than one way. Although this is not news, we will bring you up to speed in the event you have been off the grid for 12 months (lucky you).
Online shopping was becoming everyone's favourite lockdown activity (by May 2020 online orders had almost doubled in value than the year before), but supply chains were also being crippled.
The lead times for ecommerce merchants who source products and supplies from China increased from days to months. Staffing problems, new health regulations and soaring shipping prices caused bottlenecks at airports, borders and ports.
This is just one example of how sudden changes in legislation can impact supply chains. The long-awaited Brexit was much less abrupt. The long-awaited Brexit was less unexpected.
"Analysts predict that cross-border ecommerce in UK will decrease due to Brexit changes 2021," says ShipBob's Lopienski. It is therefore important to adopt a UK-specific eCommerce fulfillment strategy," states ShipBob's Lopienski.
"It is important to have a UK-specific strategy for ecommerce fulfillment."
ShipBob reports at this point that all four major carriers in the United States have met their pre-COVID average delivery time, despite persistent challenges. Future pandemics will likely occur despite the availability of vaccines. Changes to the supply chain due to political instability or natural disasters are even more likely.
ShipHero hopes that President Joe Biden will support the Buy American campaign.
Daniel-Richards says, "We expect it to bring business to US manufacturers and in doing so, supply chains can focus on eco-friendly transport options like ground freight." Although it may be more expensive initially, domestic markets can help reduce risk in the supply chain which can pay dividends over the long-term.
Returns on product
While free returns are considered a cost of doing businesses, they have also revolutionized how customers shop online. Online shoppers are more likely to order multiple sizes, colours, or products and then find the perfect fit.
Millions of returns are made each year between Thanksgiving and January, which amounts to over $100 trillion in goods. This is good customer service but can also cause problems with supply forecasting.
"The percentage of products that are returned and the reasons for them happening can vary depending on product category and seasonality," Karen Fitzgerald, senior marketing manager at Returnedly, which offers digital return experiences for direct to consumer brands.
Alex McEachern is the marketing manager at loop Returns. This app allows Shopify brands automate returns and is the most popular.
He says that many brands overlook to include returns in their inventory forecasting. It is important to know what percentage of returns can be restocked and sold again.
supplier
relationship management
Many brands overlook to include returns in their inventory forecasting.
Trends and shifting demand patterns
You can't keep up with the latest trends and fads. If you don't have enough stock, you could miss out on an influx in demand.
Ecommerce merchants that have brick-and-mortar locations may find it more difficult to manage these customers. Customers will switch channels and make it harder to predict where stock inventory will be.
Matt Warren, CEO at Veeqo, which supports ecommerce merchants with their omnichannel inventory management, says this is why more retailers are switching to an online/offline hybrid approach. Veeqo's client is a large American fashion retailer that has a large physical footprint.
Veeqo was used to transform each store into a mini-fulfillment location. This allowed them to optimize delivery times to online customers. The ability to seamlessly combine stock data with all of their online/offline sales data allows for a more accurate demand forecast. He says it's an innovative hybrid online/offline approach that commerce has been discussing for some time.
Seasonality of products
Robinson says that one of the biggest errors ecommerce merchants make when forecasting supply chain is not taking into account seasonality and current events. It's difficult to respond to holiday sales that are booming a few weeks earlier.
"Failing to factor in seasonality or current events is one the biggest errors I see... [in] supply chains forecasting."
Lead time for manufacturer or supplier
Warren was an online luxury watch retailer before he founded Veeqo. He learned from his experience that forecasting demand is only half the battle.
He says that each supplier, and sometimes each SKU, requires a different lead-time.
It is important to recognize that different products have different lead times. Also, you need to consider warehouse and shipping lead time, which can be affected by holidays overseas.
Chinese New Year can slow down fulfillments from China. Holiday peak periods may cause delays or congestion at ports, slowing delivery. It is crucial to build strong relationships with suppliers and communicate well.
Siloed data
Warren warns that siloed information can impact the accuracy of supply chain forecasting.
"Too many merchants are using different software for different aspects of their business. He says that it is difficult to work across multiple websites, marketplaces, and fulfillment locations. It's worth investing in all-in one software to unify sales and inventory data, or using spreadsheets to do the heavy lifting.
Historical data doesn't suffice
Kristjan Vilosius is the CEO and cofounder of Katana. The company offers supply management software for manufacturers and makers. He says that it's easier to make sense of events once they have occurred.
He suggests that it is better to invest in tracking and early warning systems, and find ways to make supply chain management more efficient and less dependent upon stock levels than to try to find the best forecasting techniques.
"Investing in ways that make supply chain management more efficient is often a better option than trying to find the most accurate forecasting methods."
Next steps in supply chain forecasting
There are many factors to consider when determining which forecasting method is best.
- How long do the products last? Is it possible to keep them on shelves for a long time?
- What is the average number of products sold each year?
- How do sales react to different months, seasons and sales events?
- What are the warehouse charges associated with a specific item?
- What date must you reorder inventory?
- What are your standard points for reordering?
- Are you looking for safety stock?
"Supply chain forecasting should not be a guesswork. But that's the reality of many ecommerce merchants. Daniel-Richards says online merchants must understand the impact real-time data integrations and apps could have on their inventory replenishment abilities. It's the difference in having stock or not. It's also the difference in having stale inventory, and whether you have a supply chain that works.
Comments
Post a Comment