“Push and Pull” in ABC

Those who are familiar with the manufacturing system must have heard or used this “Push and Pull’ technique. In this technique the downstream manufacturing process does not produce the part unless it is required by the upstream process. This is true for the final product also. It has got another name also “Top down and bottom up” approach. This is very popular with the planning and budgeting process. In the Top down approach the budget amount from the parent is distributed among the children with some logic. In Bottom up approach the budget amounts from various children departments are accumulated for the parent one.

In costing also people used to take every dollar from the accounting system and passed it to the final products. The accounting principles also required it to be done in that way. Later on people started using the capacity of the manufacturing equipments in the calculation of product costs. Though the machines were available for all 24 hours, their capacity was considered upon how long they are generally used, and called as ‘Normal Capacity’. The unused capacity is called ‘idle capacity’ and corresponding costs as ‘idle costs’.

Activity Based Costing (ABC) also started with the loading 100% of the costs from the resources to the cost objects through the activities. When the use of ABC was started for the services industries like banking, insurance, telecom etc. it must have started looking absurd to take all the costs of a branch of a bank to the services that it provides and then decide its profitability. This must led to the different way of putting the ABC model. I think that is when people thought of the ABC model which calculates the costs of the services based upon the volume of transactions.

The first way modeling where the 100% of costs are ‘pushed’ to the cost objects through the activities is called the ‘Push model ‘or the ‘top-down approach’ model of ABC. The other way of ABC modeling where the costs are pulled by the cost objects based upon their volume from the resources through the activities is called the ‘Pull model’ or the ‘bottom-up approach’ model. The major difference in these two models is in the push type 100% of the cost flows to the cost object. In pull model cost flown may be less or more than the cost entered in the resource module. This depends on the capacity utilization. In case of the machine we generally used the remaining capacity as ‘idle capacity’, but the in case human resources people do not like them to be called as idle. This sometimes gives a feeling that they have work but are not doing it. So we defined a term as ‘capacity available to use’. For example a branch has a ‘teller’. She will perform the cash deposit or withdrawal activities. If she does not have the sufficient number of transactions because of the footfall in the branch, then we say the rest of time she is available for other activities of the branch. She is not said to be ‘idle’.


This type of modeling has helped in calculation of product and customer profitability in the services industries very well. It has also helped in taking the costs of the internal services departments like HR, IT etc. It has been used by various consultants for a long time now and the current wave named as ‘time driven activity based costing (TDABC)’ is actually a more sophisticated version of the bottom-up or pull type of ABC modeling. This is a modeling technique and most of the commercial software solutions are capable of building such type of models. Now we will take some example and try to understand the difference.

I have learnt this concept from Greg Nolan of GJ&Nolan Co. and I will try to explain this in a similar way. Let’s take a bank is using a push approach to calculate the product profitability.

Push ABC example

The figures in the first column are from the financial accounts. The total revenue, total expenses are available and from that we can calculate the overall profitability of the bank. The revenue can be reasonably taken to the products, but the expenses are not easy to take to the products. We assume that the expenses here are taken on the basis of the # of accounts for various products. !00% of the expenses are taken to all the three products. So if you add the expenses of all the three products it matches with the total expense. With this calculation it shows that product 1 and product 2 are not doing good and product 3 is making profit.

If you ask any business manager if there anything wrong in this, invariably the answer would be ‘NO’. In absence of any cost information this would be accepted and used. The issue starts when you start getting information on a regular basis and start comparing with the earlier period(s). The changes in the figure then may not match with the changes on the business that are happening. To understand this clearly we will add a new product to this scenario and see the calculation.





With the addition of the product 4, brings its own revenue as well as expenses. Because of this the total revenue and the expenses are changed. The new amount of expenses is distributed based on the number of accounts various products. And now, the profitability of the products changes. The product 2, which was unprofitable till last period looks profitable now.

This is when the business managers start disbelieving the numbers. Quite reasonable too. Because there was no business change that happened in that period and still the product started looking profitable. The question comes, what is the correct information? The current information that the product is profitable or the earlier information that the product was unprofitable. Nobody can give the correct answer with this data.

Pull ABC Example

We will take the similar example of the products and try to understand the Pull ABC model.

What is the difference here? The expenses are not distributed based on the # of accounts, but the unit costs are calculated for each product and it is multiplied by the # of accounts. In this case the total expenses do not match with the sum of expenses of the products. This is because the balance is the capacity available for expansion.

Now we add a product here also and see the calculation.

When you add the product and the income and expenses of the same the profitability of the products 1,2 and 3 is unchanged. And which should be, isn’t it? The expenses are taken from the capacity available and the available capacity is reduced.

For the services industries most of the operating expenses are incurred in putting up the resources to service the customers and process the transaction they perform. The profit in that case is the difference between the amount paid by the customer and the amount of resource they consume, as well as the amount of resources positioned but not consumed.

We will take another example of a withdrawal from an ATM. There is an ATM in remote corner of the city, which is hardly used by any customer, but put up by the bank. One fine day a customer goes to that ATM and withdraws some money. In case of the Push model, the whole of the quarter’s expenses for the ATM would be charged to customer and the poor customer would not be shown profitable for the entire life with the bank.

In case of Pull model, the cost of one transaction would be calculated based on the capacity of the ATM to handle and time taken by the transaction. Based on this calculation the customer would be charged to the customer.

The Pull modeling technique can also be used for resource planning. Based on the volume of the different products to be sold in future, we can calculate the requirement of various resources like machines, skill set etc. by adding budgeted expenses to this model we can calculate the planned profitability of the organization. Using this planning model we can compare the actual expenses, profitability, utilization etc. and understand the reasons for the variance.