Implementing Activity Based Costing ABC/M without tears
(PART 2)
10th September 2007 In his last article, John McKenzie, FSN contributing editor, examined the first steps of a methodology for implementing ABC/M – Process Mapping and constructing an Activity Dictionary. Here he continues the implementation methodology, looking at the benefits of facilitated workshops for process mapping and other practical techniques for identifying activities, processes and cost drivers without overwhelming the organisation.
The most effective way of constructing an activity dictionary is through multi-functional workshops, each consisting of 7-8 people from functions with the relevant knowledge and experience of front line operations, along with a facilitator to guide the whole process. It's tempting for senior management to get involved in such workshops, but in reality, few have or are expected to have the level of detailed knowledge of operations to contribute effectively. Junior management, supervisors, team leaders and ordinary staffs are more likely to achieve results. A workshop usually lasts a day and focus on 3-4 processes, grouped by similarity of involved functions. Process mapping of an organisation or division or business unit might reveal in the order of 15-20 meaningful business processes. As a result 5-6 workshops will be conducted but is time well spent as it greatly simplifies subsequent data collection and ensures a common approach. It also creates “buy-in” amongst staffs who see the ABC/M project as more involving rather than imposed.
The facilitator's role is vital, firstly to ensure that the number of activities is limited to that needed for the output desired as discussed previously. Additionally, the facilitator would lead participants to identify what drivers are relevant for each activity. Finally the facilitator directs the workshop team to focus upon activities related to process failure. Processes are rarely designed to contain non-value added activity from the outset. Most are caused by conditions created either internally, or by supplier or customer actions that result in a need to deviate from the ideal process and into wasteful “error correction” loops. An example might be debt chasing – contentious perhaps, but we will see why it is non value adding in a later article covering value analysis and re-engineering. Further wasteful activity can also be identified in areas of duplication of activity and tasks, over-control or bureaucracy, inspection and checking and delay time.
A sample of a dictionary would look like this
Activity
Value Added
Time cycle
Driver
Enter Customer Order to System
Customer Value Adding
Continuous
# of orders
Run Order Processor
Customer Value Adding
Daily
# of orders
Identify Out of Stock Items
Non Value Adding
Daily
# of stockouts
Chase Supply chain for Out of Stock Items
Non Value Adding
Continuous
# of stockouts
Produce and Send Order Confirmation
Customer Value Adding
Daily
# of orders
Generate Picking Lists
Customer Value Adding
Continuous
# of orders
Pick Customer Order
Customer Value Adding
Continuous
# of orderlines
Pack Customer Order
Customer Value Adding
Continuous
# of orderlines
Dispatch Customer Order
Customer Value Adding
Continuous
# of cases
Invoice Customer
Customer Value Adding
Continuous
# of orders
Collect and Bank Payment
Customer Value Adding
Continuous
# of orders
Chase Overdue Payment
Non Value Adding
Continuous
# of overdues
Construct Annual Budget
Sustaining
Annual
--------
Review Monthly Budget Performance
Sustaining
Monthly
--------
Attend Weekly Management Meeting
Business Value Added
Weekly
--------
3. Ledger and Organisational Analysis
This is a designed to achieve 3 things:
First, to ensure the costs as held in the ledger are in line with the organisational structure. Over time these can sometimes go “out of sync” due to structural changes or transfer of resources. If there are discrepancies these have to be dealt with either within the ledger or by adjustment in the ABC/M model itself to bring them in line.
Second, the character of cost held in the ledger must be identified and the allocation method selected to apportion these costs to activity. This process was described in the first article of this series. Specific items such as depreciation and travel costs will need a more detailed analysis to establish apportionment.
Finally, the analysis should identify managers, team leaders or supervisors who will be called upon to participate in the data collection process in steps 4 and 5.
4. Management Briefing and Communication
The most arduous part of an ABC/M project is the collection of data relating to how people spend their time and hence the allocation of employment and many related costs. This is usually done on a percentage of time spent basis. There are 3 main ways to collect time data.
The first is by direct observation or time study. This is only feasible where work performed is of a highly repetitive nature, non-cyclical and the activities performed in the work unit being examined are very limited in number.... for example, production line operations, call centres, etc. The danger of direct observation is that people may cease to behave “normally” whilst under observation and if there are activities that are cyclical or infrequent, these may not be captured in the observation period.
The second method is by time sheets. These generally fail to work satisfactorily unless there is already a well established and disciplined time sheet recording process. Time sheets used for “one-off” or infrequent observations are frequently misunderstood or misinterpreted, open to abuse and have been shown to be far more inaccurate than the third option which is survey and interview using the standard activity dictionary developed above. Individuals identified in step 3 should be brought together for a briefing covering a high level understanding of ABC/M, the purpose of the project, what an activity dictionary is for and their role in the project. At the end of these sessions (there may have to be several to cover all the individuals without bringing work to a standstill), all participants will be given a copy of the dictionary to take away and perform Step 5.
5. Data Collection and Interviewing
Using the activity dictionary as a starting point, each participant is tasked to conduct a time review for themselves and the staffs for which they are directly responsible. This is achieved by the participant working together with their staffs in discussion to identify which activities in the dictionary apply to the work unit and then to consider as individuals and collectively how work time is spread across them. A time analysis return might look like this:
In this example (the analysis of an Order Management department) there is 10 staff of differing seniority/grades, separated out as the employment costs of each differ due to pay grades and attracted benefits. It has been determined that of the activities contained in the dictionary, those listed above apply to this department. Each individual has 100% of time available (therefore with 10 in this department there is 1000% of total time available) and this is spread across the activities. Attempting to account for time down to the nearest minute is spurious. Generally, accounting for time to the nearest 3 to 5% by activity is accurate enough. The column referring to weighed average is a computed column based on the time spent by staff category and the cost of each category and represents how the overall department employment costs should be spread.
Collecting time data in this way utilises a small, manageable amount of time of a large group of people and if co-ordinated well means that even thousands of staff can be analysed in a rapid timescale – typically as little as two or three weeks. Most submissions will be satisfactory but it may be necessary to conduct a number of follow up interviews. For example, in the case of complex work units or where the submission has activity times that seem unusual compared to comparable activities performed elsewhere.
6. Driver Collection
A raw output from the dictionary workshop might identify 100 or more drivers…. a nightmare task for data collection. In reality 30-50 are likely to have a significant impact or materially affect the analysis outcome. The trick is to identify which 30-50. You have to consider which drivers are vital and those which are likely to drive minimal costs for which an acceptable proxy can be drawn from the vital ones. Driver data has to be collected in some way and most of the major ones may already be captured in existing ERP, call management, financial or transactional systems, limiting the task to extraction from these sources in a format ready to load into an ABC model. A few however may be currently unmeasured and if these are important enough to affect materiality of the outcome then there is no option but to either perform a direct observation or count, or to undertake a manual analysis of offline reports or paperwork if such sources exist.
7 Model Building
Software issues and related modelling techniques will be dealt with at the end of this series of articles. Suffice to say at this point that if you are modelling at the strategic level or undertaking a proof of concept exercise, it is perfectly practical to build a robust and satisfactory model in a spreadsheet tool such as Excel. As modelling complexity rises when working at the tactical level then due consideration must be given to the acquisition of a specialist modelling tool of which there are several available.
The key issues with modelling are to keep it simple as possible to aid long term maintenance and before even touching a modelling tool, draw the cost flows out on paper and identify the modelling logic. This may seem a retrograde step but its value is enormous in preventing the construction of a model only to find mid way through that the underlying logic is flawed and starting over again.
In the next article we begin to look at the applications of ABC/M, starting with its traditional use for profitability analysis of customers, products and distribution channels.
