New Seller Incentives: Over $50,000 in Referral Fee Savings per MarketplaceNew Seller Incentives: Over $50,000 in Referral Fee Savings per Marketplace
Should cross-border e-commerce sellers evaluate the accuracy of sales forecasting
preface
Many sellers immediately think that this is a problem with sales forecasting due to inventory backlogs or shortages.
An assessment is then required to assess the accuracy of sales forecasting.
Then someone asked me how to evaluate the accuracy of sales forecasting.
Today's topic, let's talk about this.
Many sellers immediately think that this is a problem with sales forecasting due to inventory backlogs or shortages.
An assessment is then required to assess the accuracy of sales forecasting.
Then someone asked me how to evaluate the accuracy of sales forecasting.
Today's topic, let's talk about this.
1. Inventory issues. First, we must eliminate the problem of stocking logic
As soon as everyone talks about inventory issues, the first thing that comes to mind is the issue of sales forecasting.
It's not like that.
The first thing to eliminate is the problem with the stocking logic. If there is a problem with your stocking logic itself, no matter how good the sales forecast is, there will also be a problem of out of stock or backlog.
The stocking logic involves many aspects. For example, we need to consider our inventory not only from the perspective of sales volume, but also from the perspective of Amazon's policies. For example, the inventory level in FBA must not be too low; otherwise, there will be low inventory level fees. (See Low Inventory Level Fee), in addition, high-volume ASINs recommend small, frequent restocking: every three weeks, as it is necessary to consider Amazon FBA's internal cargo transfer logic and distribute the goods as close as possible to the operating centers closer to the consumers.
Then there are many aspects to rule out implementation issues. Let's just talk about procurement. If your buyer is not binding on the supplier, most of the supplier's delivery is delayed, sometimes by 3 days, sometimes by 6 days, sometimes by more than 20 days, and supply stability is completely uncontrollable. Under these circumstances, it is difficult to accurately calculate the water level. Give more, stock more; give less, and seriously run out of stock. Therefore, in this case, accurate sales forecasting is useless.
After excluding the logic of preparation, stability of procurement, and stability of logistics, let's talk about sales forecasting.
It's not like that.
The first thing to eliminate is the problem with the stocking logic. If there is a problem with your stocking logic itself, no matter how good the sales forecast is, there will also be a problem of out of stock or backlog.
The stocking logic involves many aspects. For example, we need to consider our inventory not only from the perspective of sales volume, but also from the perspective of Amazon's policies. For example, the inventory level in FBA must not be too low; otherwise, there will be low inventory level fees. (See Low Inventory Level Fee), in addition, high-volume ASINs recommend small, frequent restocking: every three weeks, as it is necessary to consider Amazon FBA's internal cargo transfer logic and distribute the goods as close as possible to the operating centers closer to the consumers.
Then there are many aspects to rule out implementation issues. Let's just talk about procurement. If your buyer is not binding on the supplier, most of the supplier's delivery is delayed, sometimes by 3 days, sometimes by 6 days, sometimes by more than 20 days, and supply stability is completely uncontrollable. Under these circumstances, it is difficult to accurately calculate the water level. Give more, stock more; give less, and seriously run out of stock. Therefore, in this case, accurate sales forecasting is useless.
After excluding the logic of preparation, stability of procurement, and stability of logistics, let's talk about sales forecasting.
2. Sales forecasting cannot be accurate
Sales forecasting cannot be accurate. This can be considered the primary principle of supply chain planning, which is particularly prominent for cross-border e-commerce.
The reason is simple; let's just talk about competitive factors: in a competitive environment, we can never predict how so many competitors will act.
Simply put, most of the time, predictions are accurate; it's not that you've made good predictions, but that the external environment hasn't changed much.
In other words, if the accuracy of sales forecasting must be assessed, then this kind of assessment is meaningful when there is no link to changes in the external environment. It will guide you to optimize parameters and do better.
There is no point in evaluating links that have clearly changed in the external environment.
Also, speaking of accuracy, what is accuracy?No one more, no one more, no one can do it. The accuracy of what is stated here is within the acceptable range of deviations. For example, within + -20%.
The reason is simple; let's just talk about competitive factors: in a competitive environment, we can never predict how so many competitors will act.
Simply put, most of the time, predictions are accurate; it's not that you've made good predictions, but that the external environment hasn't changed much.
In other words, if the accuracy of sales forecasting must be assessed, then this kind of assessment is meaningful when there is no link to changes in the external environment. It will guide you to optimize parameters and do better.
There is no point in evaluating links that have clearly changed in the external environment.
Also, speaking of accuracy, what is accuracy?No one more, no one more, no one can do it. The accuracy of what is stated here is within the acceptable range of deviations. For example, within + -20%.
3. Sales forecasting is a process indicator, not an outcome indicator
Assessing sales forecasts requires a clear understanding: sales forecasting is a process indicator.
Process metrics are good, but this might not be the results you want.
A company assessed the accuracy of sales forecasting. As a result, sales were very tight. As a result, the forecast was very conservative, and it was impossible to let go even if it was ranked.
For example, this product can do: 300, 500, 900, 100 in the next 4 months.
Because of the assessment, sales forecasts are conservative: 200, 400, 700, 50.
What was the result? The sales volume was as predicted.
However, it was possible to get higher rankings and more sales, but the boss was supposed to earn money, but he didn't earn it.
Is this the outcome you want, boss?
Another opposite example: if the process indicators are bad, the results may not be bad.
For example: Sales forecast a product with a strong season in the future. This product can do it in the next 4 months: 300, 500, 900, 100.
As a result, just a month passed, and only 180 were achieved.
The products have already been purchased and sold through a period of price reduction, promotion, etc., and finally the inventory was sold out at the end of 4 months.
In this sense, the process indicators are not good, but the inventory results indicators are good (of course, not necessarily good from a profit perspective).
Process metrics are good, but this might not be the results you want.
A company assessed the accuracy of sales forecasting. As a result, sales were very tight. As a result, the forecast was very conservative, and it was impossible to let go even if it was ranked.
For example, this product can do: 300, 500, 900, 100 in the next 4 months.
Because of the assessment, sales forecasts are conservative: 200, 400, 700, 50.
What was the result? The sales volume was as predicted.
However, it was possible to get higher rankings and more sales, but the boss was supposed to earn money, but he didn't earn it.
Is this the outcome you want, boss?
Another opposite example: if the process indicators are bad, the results may not be bad.
For example: Sales forecast a product with a strong season in the future. This product can do it in the next 4 months: 300, 500, 900, 100.
As a result, just a month passed, and only 180 were achieved.
The products have already been purchased and sold through a period of price reduction, promotion, etc., and finally the inventory was sold out at the end of 4 months.
In this sense, the process indicators are not good, but the inventory results indicators are good (of course, not necessarily good from a profit perspective).
4. The results of sales forecasting cannot be directly used to prepare goods
Another point is that it's probably easy for everyone to make mistakes: sales forecasts are used directly to prepare goods.
In fact, this is not the case; there is an important part in this, the safety inventory setting.
Cross-border e-commerce safety inventory includes two aspects: one is safety inventory to deal with unstable supply, and the other is safety inventory to deal with increased or fluctuating sales.
The traditional supply chain theory is only safe inventory with fluctuating sales volume, but in e-commerce platforms, the focus is on safety inventory with increased sales volume.
On the one hand, stored safety inventory can meet the needs of increased sales; on the other hand, it also enables FBA inventory to meet minimum inventory levels.
After processing, the sales forecast becomes a demand plan, and the calculation of inventory volume is based on the demand plan.
In a nutshell, even if the sales forecast is good, it's useless if the stocking plan isn't done well.
In fact, this is not the case; there is an important part in this, the safety inventory setting.
Cross-border e-commerce safety inventory includes two aspects: one is safety inventory to deal with unstable supply, and the other is safety inventory to deal with increased or fluctuating sales.
The traditional supply chain theory is only safe inventory with fluctuating sales volume, but in e-commerce platforms, the focus is on safety inventory with increased sales volume.
On the one hand, stored safety inventory can meet the needs of increased sales; on the other hand, it also enables FBA inventory to meet minimum inventory levels.
After processing, the sales forecast becomes a demand plan, and the calculation of inventory volume is based on the demand plan.
In a nutshell, even if the sales forecast is good, it's useless if the stocking plan isn't done well.
5. How to assess the accuracy of sales forecasting
The traffic mechanism, sales principles, and competitive trends of cross-border e-commerce are very different from traditional offline stores.
The accuracy index of sales forecasting, if assessed:
1) Suggest a planner for baseline forecasting (first you have to have a planner, then a planner who specializes in baseline forecasting);
2) The purpose of the assessment is to urge baseline forecasting planners to improve and improve the accuracy of baseline forecasting;
3) The scope of the assessment is: for those links -- links with a stable external environment in hindsight;
Of course, there are many details to consider. For example, at the beginning of May, planners predicted sales for May, June, July, and August.
Let's say we want to evaluate the accuracy of the sales forecast for month 8.
If you make a monthly baseline forecast, there are at least 4 rounds of forecasting sales for August, at the beginning of May, the beginning of June, the beginning of July, and the beginning of August, respectively.
If a baseline forecast is made every 2 weeks, then there are 8 rounds of predictions;
If the baseline forecast is made once a week, then there are 16 rounds of predictions;
What round of predictions will it be used to evaluate it?
The accuracy index of sales forecasting, if assessed:
1) Suggest a planner for baseline forecasting (first you have to have a planner, then a planner who specializes in baseline forecasting);
2) The purpose of the assessment is to urge baseline forecasting planners to improve and improve the accuracy of baseline forecasting;
3) The scope of the assessment is: for those links -- links with a stable external environment in hindsight;
Of course, there are many details to consider. For example, at the beginning of May, planners predicted sales for May, June, July, and August.
Let's say we want to evaluate the accuracy of the sales forecast for month 8.
If you make a monthly baseline forecast, there are at least 4 rounds of forecasting sales for August, at the beginning of May, the beginning of June, the beginning of July, and the beginning of August, respectively.
If a baseline forecast is made every 2 weeks, then there are 8 rounds of predictions;
If the baseline forecast is made once a week, then there are 16 rounds of predictions;
What round of predictions will it be used to evaluate it?
Having said so much, let me summarize my personal opinion:
To assess the inventory performance of sales, this awareness is correct; sales must have the motivation to control inventory.
However, the assessment method should start with the results. The result we're talking about here is inventory — whether it's out of stock, whether there's a backlog.
A good result means:
1) Keen insight into products and external environments, able to discover opportunities or risks;
2) After discovering opportunities or risks, whether it is later operating methods or preparation, there is an accurate and quick adjustment strategy;
3) These strategies can be implemented effectively;
4) Get the results in the end;
Of course, this result doesn't have to be calculated after half a year or a year; it has to be calculated every month.
To assess the inventory performance of sales, this awareness is correct; sales must have the motivation to control inventory.
However, the assessment method should start with the results. The result we're talking about here is inventory — whether it's out of stock, whether there's a backlog.
A good result means:
1) Keen insight into products and external environments, able to discover opportunities or risks;
2) After discovering opportunities or risks, whether it is later operating methods or preparation, there is an accurate and quick adjustment strategy;
3) These strategies can be implemented effectively;
4) Get the results in the end;
Of course, this result doesn't have to be calculated after half a year or a year; it has to be calculated every month.
The above content only represents the creators' personal opinions. The data is for reference only, and does not represent the official views of Amazon Global Store.
Getting Started: New Seller Guide
Global business opportunities
