What you plan for tommorow you shouldn’t give away today…
Everybody knows how a VSF generated by strategy 40 works, right? After all its the most commonly used planning strategy around.
You enter a forecast (there are many automated or manual ways to do it) and the MRP planning run ensures that there are supply elements to provide enough product so that if the customer wants what you forecasted, everything will be fine. Of course, the forecast is never the same as sales are, but that is what safety stock is for.
So if your planners are somewhat right, your MRP controllers make sure that product is made and available just before the forecast requested it and your sales people check availability on available stock, You will be able to have a very high fill rate.
But there is another component you have to worry about; the way you consume your forecast!
The consumption strategy you find on the MRP3 screen. There is the possibility of consuming forward, backward or both. And not too long ago I thought going backward first and then forward is the best thing to do. I don’t think that way anymore…
Just recently a production scheduler complained to me that he can’t handle all the requirements thrown at him… Ever!
We looked at many things. Availability checking rules, forecast accuracy, demand leveling, available capacity… But the problem, as it turned out, was the consumption of the forecast.
What happened was that there was strategy 40 (a make to stock strategy where a forecast is consumed until its gone and then additional demand is placed on the line) and a consumption strategy that looked 20 days backwards first and then 22 (working) days forward to see if there is any forecast that can be used up.
Now the product we were looking at was an A item and a big seller which was chronically under-forecasted. My friend (he is now a good one 😉 ) was complaining that no one ever requested anything in advance, but the orders rolled in last minute in quantities impossible to handle.
Well… Think about it… Too low of a forecast and then a consumption logic where sales orders quickly run the current forecast down and then consume down the forecast of the next month and then the one thereafter. And if there is no more forecast left for the next month, the planning run thinks you don’t need anything the next month.
So the more you sell the less is planned for. Not a good planning strategy!
Here is a simple remedy: consume ONLY backwards. If you do the right thing, there will be inventory on the first day the forecast stands at and throughout the period you consume THAT forecast. After you used it up you have safety stock. And after that is gone, you can tell my production planner friend to make more – but only then – and that will be the exception and most likely at the end of the period. But don’t ever take down the forecast of the next period; it does not make sense to tell the system that if I sell a lot today, I need less tomorrow.
???? Surely you are joking. Reducing forward consumption is a simplistic answer that is usable only in very limited circumstances.
In most companies, the major goals are to maximize sales revenue, while minimizing inventory build. If the sales orders have been confirmed, then it implies that the excess sales requirement over forecast has been deemed as 'manageable'. Changing the forecast consumption won't change that business practice. If there is an unsustainable mismatch between forecast and actual, and that mismatch causes unforecasted perturbations in the actual manufacturing plans, then the solution is a change in ATP policies, not a change in forecast consumption.
Schedulers will always complain, that is their role in life. That complaint by itself is not an adequate reason to change any business practice.
On major reason Forward and Backward consumption are used is to accommodate the situation when sales demand is 'lumpy', and the forecast is 'smooth'. As in 'just about always'. If you reduce or eliminate forward consumption, then you increase the risk of building inventory that will never be sold. In many companies, this would be grounds for dismissal.
Uwe,
I am not buying your defense of the schedulers. In the end, the company has decided that It will take the business, and the schedulers must find a way to meet the demand derived from actual sales. That's why they get a paycheck.
It is a natural tendency for marketing/sales to underforecast, especially if they are not held to any standards (measurables). The solution to this problem is not to change the consumption of the forecast, it is to correct the forecast. All forecasts are 'wrong' in some way; but ALWAYS being wrong on the short side is bad for the company (increases manufacturing costs), and this metric can be improved - provided that there is motivation to improve.
Reducing the forward consumption means that you are going to build next month's forecast without regard to the likelihood that sales orders will/will not be entered next month. Dangerous stuff.
You set the forward consumption after you have analyzed the quality of the forecast, and the variability of the actual sales. Arbitrarily reducing consumption to zero to protect the 'poor' planners is short sighted indeed.
I hasten to add that among the many hats I have worn during my business career were those of both a planner and a planning manager.
I don't disagree. What can be sold should be met by production! No question there. However, in my mind, not the forecast should be adjusted but the plan. A forecadt is a forecast is always wrong. And if you adjust the forecast you cover up the problems. I usually tell my clients to perform an ABC / XYZ analyse to first determine what CAN be forecasted. Then they can put a forecast out there to which the production planners can schedule the line... But only to 75% capacity! That way there is room to be agile and react quickly to additional sales.
But when you adjust the forecast its not a forecast anymore and you cant say how good or bad your planning was.
By the way... I wholeheartedly agree with you that the ATP checking rules are way more important... And there are many other factors too. This blog post wanted to point out a specific problem I had at a specific client. I am hoping that this kind of thinking might help others too...
To sum it up: i believe that when you consume the forecast forward and you have more sales in the current period than you forecasted for the current period, you need to adjust the production program of the current period - not the forecast of the future period.
Uwe,
I should not have been so vague in my answer, my apologies. The forecast method is what must be changed.
If the forecasting method produces forecasts that are smaller than actual 50% of the time; and are greater than actual 50% of the time, then one could make the argument that the forecasting method was statistically valid. 'Chronic' underforecasting should be analyzed to find the root cause. It is almost as bad for the company's bottom line as chronic overforecasting.
This is seldom necessary. The manufacturing managers have already told huge lies to sales and to the planners about what the actual manufacturing capacity is. The capacity they gave to you was already 75% of maximum.
To sum up my perspective on your summary, yes, there are valid reasons to reduce the consumption periods, both forward and backward. In my opinion, using this solution as a workaround for crappy forecasting methods is not one of them. Your enthusiasm for this solution, which obviously solved one problem for one client (e.g. it made the schedulers happier), should not be taken by the readers of this blog as an indicator that it is a cure-all for any other problems.
I must say you have an interesting view on things... and they are very dogmatic to say the least... so the production planners have already lied to me, huh? and it's seldom necessary to plan to 75%??
ok, so the forecast method is what MUST change? since when should people adjust a prediction to which we execute (that is what make to stock is: executing to a forecast)
and since you already know what I do at my clients (to make the production schedulers happy), please let at least the readers take this blog whichever way they want... you did that too
@ Uwe
Some thoughts:
If forward consumption is not allowed, but sales quantity in present month adds into existing PIR demand (w/o consuming future PIR within a defined window by business), then wouldn't it disturb the entire production planning?
Also the proposal made in this blog might work in some situations but definitely not in all. A lot of factors need to be kept in mind, for eg: the product life cycle stage in which that part / product is in etc.
Hi Vivek
I agree wholeheartedly with your second statement. this blog post addresses a specific problem (where the forecast was always too low, and therefore a reduction of future forecasts made the problem worse)
in regards to your first statement, i would like to say that if you forecast to sell 5000 pieces in June and forecast 5000 pieces in July, the production planners can plan to make these quantities available. Should actual sales in June exceed the forecast, additional requirements should be introduced into the June production program - rather than consuming the July forecast and making everything look like it is ok. Noise is introduced to the plan in either case, but why reducing the predicted sales in the future? doesn't that make things worse?
A demand planner periodically checks the consumption / reduction of the PIR's placed & by not allowing forward consumption, the planner will have to do the math outside of SAP or a Z report to see by how much the PIR's fell short vis.a.vis sales.
Yes any change always tags along disturbance with it. But by not allowing forward consumption, that disturbance is brought forward & the entire supply chain (internal & external) will be under stress.
Anyway all solutions needs to be in context, maybe in the context which you've implemented that works fine.
yes, I agree. I guess my point is, that I don't believe that when you forecast 5000 for June and 500 for July, the actual sales in June should only reduce the June forecast and not the July forecast. So if the June sales exceed the June forecast, then safety stock shuld be used and July forecast left alone. If it even exceeds safety stock, it should place additional demand.
i just don't understand why more sales in june should bring down the July forecast and therefore result in lower inventories in July to pick from. So if the June forecast reduces the July forecast down to 4000, does that mean that we sell only 4000 in July??? thats what does not make sense to me, but maybe I am wrong and both you and dogboy are right... almost for sure forward consumption has its validity in certain cases, but my client was chronically underforecasting and the forward consumption cut their future orders for upply in half... and that was disastrous
?? shrug...??
Uwe,
<< shrug >> from me too.
In the end, if the client has a smile on his face, and the invoice gets paid, then the solution you gave him was the right one. 😉