Sanac Inc.: From ABC to Time-Driven ABC 1.1 Introduction We are competing in an environment with a wide variety of customers, many different products and complex services provided to the customers. Our sales are growing, as is operational and administrative complexity, but profitability is suffering. We urgently need detailed insight into customer profitability. Although ABC seems like the perfect textbook tool to understand costs and profits, in this situation given the increased complexity of the business and our seasonal demand, it turns out to provide inaccurate information at best. How can that be and what can be done about it?--- Gertjan De Creus, CEO of Sanac. In 2000, Gertjan De Creus was appointed as the new CEO of Sanac Company, located in Wervik, Belgium. His mission was to change the company culture from revenue to profit maxi-mization. To retain customers and to defend market position in a fiercely competitive market, sales reps at Sanac were promising more services to customers. As a result, the company was growing, but profits were falling. De Creus wanted to have detailed monthly reports on customer profitability. In June 2005, the company started an ABC project. After reviewing the initial re-sults, De Creus doubted whether ABC could ever be successfully implemented in his company. 1.2 Company Background Sanac is a family-owned distributor of plant-care products in Belgium; the company posts total sales of 62 million, a transportation fleet of 25 trucks, and a warehouse of 22,500 pallets of products. Sanac neither produces nor sells any of its products under its own name. It only retails well-known brand products to three different customer groups: farmers, growers, and retail out-lets (large and small department stores, do-it-yourself shops, and garden centers), as shown in Table 1. Sanac has about 7,000 customers, 7,000 products, and receives approximately 298,000 order-lines per year. The company employs 129 people, 40 of whom are sales reps and 57 work in the warehouse. The company has been in continuous expansion; in the last 10 years its sales have increased by 10 per cent annually. Sanacs competitive strategy is predicated on three core elements. The first is product know-how. Sanac not only sells items related to planting and growing (especially the plant-care products) but also distributes its product knowledge to the customer. Sanac compares itself to a pharmacist: its mission is to help farmers and growers find an appropriate phyto and fertilizer. Based on this product know-how, Sanac builds long-term relationships with farmers and growers. For the retail outlets, Sanac compares itself more to a consultant, by providing displays to the out-let stores. The purpose of a display is to provide information to the final consumer so that (s)he immediately finds the right product for typical garden problems (e.g., insects on roses). Often, employees of retail outlets do not have the appropriate skills to guide the final consumer. Sanac, in fact, operates a sales area within the retail outlet, ensuring its profitability per square meter, without any, or very little, investment or training required on the part of the outlet store. The second key component in Sanacs strategy is market know-how. Sanacs markets are seasonally sensitive and subject to trends. Therefore, feedback from sales technicians and sales representatives is essential as this feedback constitutes an important source of information on what is happening in the market. Also, information from truck drivers is important because these individuals know the customer very well. Trend analysis, visits to fairs, and trend-watching are important to detect trends for the next season. If pottery needs to be procured in China, Sanac must know what kind of color customers will prefer in the coming year. This market knowledge is important to avoid unsold items. Third, providing services to the customer is important. Sanac is a logistics company; fast and on-time delivery are important since the company deals either with professionals within su-permarkets and hypermarkets (for whom any delay means a loss of sales) or with farmers and growers (for whom speed-of-delivery may adversely affect their production). 1.3 Profitability Paradox In the 1990s, Sanac was primarily sales driven. There was a total lack of insight into the relative profitability of customer segments. In the 2000s, competition increased, profit margins came under pressure, and customers asked for more services. Furthermore, the company faced a fast-changing market, strong seasonality, growing complexity, and diversity of its products and customers. These changes forced Sanac to switch from a growth strategy to a profitability-enhancing strategy. Marketing strategy was focused on the retail outlets (the so-called home & garden busi-ness), which was a new segment since the early 2000s. However, Sanac did not understand the true cost of this new activity or the resources invested in it (e.g., the logistics, administration, and financing costdue to long payment terms). Despite a significant increase in annual company sales (from 37 to 62 million in 4 years) and stable gross margins, a decrease in operating profit was noticed. To address this paradox at Sanac, Gertjan De Creus wanted more insight into the cost data. The existing cost accounting system only allowed for the calculation of a contribu-tionmargin (i.e., sales minus cost of purchases) for major product lines. (See Table 2.) There remained, however, unallocated overhead (approximately 2.5 million in 2000) and these costs, considered fixed in principle, were actually increasing each year. Based on the annual increase in overhead costs, and the fact that customers did not have to pay for the increased ser-vices (e.g. delivery terms), De Creus believed that the home and garden business was generating significant losses, that agriculture was profitable, and that horticulture was slightly loss-making. In fact, the line of business with the highest growth potential was probably also the least profita-ble line. Sanac was changing its business but the financial complexities associated with the new business were not well understood. Table 1 Lines of Business (LOB)Characteristics Agriculture Horticulture Home and garden Activities Purchasing Purchasing Purchasing Technical support Technical support Product management Marketing Marketing Marketing Sales Sales Sales Customers Farmers: small to large Growers: small to large Garden centers: small to large DIY stores: small to large Department stores (e.g., Hubo, Makro, Gamma, Leroy Merlin, Carrefour) Products Professional products Products for individual consumers Phyto (crop protection) Phyto Phyto in small packs Fertilizer Horticultural supplies Fertilizer in small packs Seeds (e.g., small pots) Garden furniture Other Cutting tools Pottery and deco Garden consumables e.g., swimming pools Sales Strong seasonal trend Strong seasonal trend Strong seasonal trend Market stage Declining market Stagnating market Growing market Competitors Many Many Strong position of Sanac Table 2 Contribution margin in 2005 Line of Business (LOB) Contribution Margin Increase in Sales Agriculture 17% + 3% Horticulture 20% +10% Home and Garden 24% +20% 1.4 Activity-based costing (ABC) The reduced profitability (20 per cent decrease in the annual profit to sales ratio over the period 2000 to 2004; almost 100 per cent increase in sales from 2000 to 2004) meant that a more accurate cost system was needed. Sanac changed its focus from growth to profitability. The growth focus forced the company to have a wide range of products, to maintain large stocks, and to deliver any order to any customer in the shortest possible time in whatever quantity was re-quested. Further, customers did not pay a specified premium for the costs associated with the ser-vice level provided. Gertjan De Creus commented: We apparently have good and bad customers. Our good customers generate margins and profit, but we do not know who they are. If we do nothing, we will end up with only bad customers. I need cost information per business, per customer, per order, per delivery, per order-line, per product, and per supplier. All of this is needed to provide profitability insight. Afterwards, we can start talking about whether to discard certain clients, products, suppliers or even businesses. But for this to occur, or for any change in strategy, we first need a reliable system for measuring costs and profit. In June 2005, Gertjan De Creus directed his controller, Mike Johnson, to implement an ABC system. Johnson hired a leading consultancy company to help him design the ABC model. To start the process, consultant Chris Miller interviewed the heads of the functional departments about the activities they performed and the percentage of time they spent on the various activities. Next, Miller designed an activity database and identified appropriate activity cost-drivers. How-ever, it soon dawned on him that a traditional ABC system would represent a huge challenge for Sanac. Chris Miller had a discussion with Gertjan De Creus about the requirements of the ABC model. Gertjan De Creus: I absolutely need monthly profitability reports at the level of business, customers, orders, products and suppliers. Chris Miller replied: Can you imagine how much the ABC system will cost you to provide monthly profitabil-ity reports based on actual sales and services provided? I guess Sanac will need to hire 10 controllers to update the model monthly. But Gertjan De Creus progressed: The model should be updated monthly with up-to-the-minute data. The rapidly growing market of retail outlets requires me to act immediately on a customer request. Outlet stores are very demanding when it comes to price and services. Sanac has to be an agile company, able to quickly change processes. In such a dynamic environment the adjust-ments of the ABC model to changing service requests of customers is very important. Miller looked at the data and explained: The great complexity and diversity of customers and products means that there are just too many activities to keep track of. Sanac has 7,000 clients, a portfolio of 7,000 products in stock and 20,000 in catalogue. It receives approximately 298,000 order-lines a year and issues 69,000 invoices. But more importantly, clients have varied demands in terms of services. This wide diversity in the consumption of resources makes it difficult to trace and analyze costs. There is a wide discrepancy in resource consumption according to products (e.g., type of product, type of packaging) and customer behavior (e.g., order fre-quency, delivery terms, visits of sales technicians or sales rep, payment terms). Of course, a small farmer does not generate the same profit margin or the same costs as the outlet store of a large chain, but how can I accurately include in the ABC model their relative service demands? Chris Miller looked further into the activity cost-driver selection. The strong seasonal trend in sales made it difficult to define a normal capacity for each of the activity cost-drivers. About 80% of the sales were realized in four (summer) months; there was significant idle capacity in the winter months. It was clear that Gertjan De Creus wanted to know each month which departments were faced with overcapacity and which were running under capacity. Based on the data of the first six months of 2005, Chris Miller assigned the costs to the individual activities. For the activities sales-order processing and picking a delivery the following information was collected (see Table 3). Based on this information the activity-cost driver rates were calculat-ed, as shown in Panel A. In December 2005, Mike Johnson recalculated the activity cost-driver rates. He entered into the ABC system the total costs and total driver volumes over the last 6 months, as shown in Panel B of Table 3. De Creus observed a significant change in the activity cost-driver rates and asked Mike Johnson to check whether the cost assignments were made correctly. Mike explained that the higher activity cost-driver rates were due to the low season. But De Creus could not un-derstand the logic. He could not understand why customers were apparently using more resources in the off-peak season than in the high season. Mike Johnson explained that he had used the actu-al number of orders and the actual number of deliveries to calculate the activity cost-driver rates: I could also use practical capacity for the orders and deliveries, i.e. the largest number of orders (or deliveries) that could be handled without creating unusual delays or forcing overtime work. Using such capacity levels will result in activity-cost driver rates that stay the same during high and low season. During the low season, we could then calculate the unused capacity. But our people keep saying that it is almost im-possible to estimate what is called the practical volume of orders or practical volume of deliveries that they can handle without forcing overtime work, because the practical volume really depends on the characteristics of the orders and deliveries to handle. That is why I used the actual volume of the activity-cost drivers. Table 3 Average monthly cost data for sales-order processing and picking a delivery* Panel A: January June 2005 Activity Number of FTEs Total Cost Activity Cost-Driver Driver volume Activity Cost-Driver rate Picking a delivery 7.5 177,840 Number of deliveries 30,000 5.93 per delivery Sales-order processing 5 106,704 Number of orders 25,000 4.27 per order Panel B: July December 2005 Activity Number of FTEs Total Cost Activity Cost-Driver Driver volume Activity Cost-Driver rate Picking a delivery 7.5 177,840 Number of deliveries 12,000 14.82 per delivery Sales-order processing 5 106,704 Number of orders 10,000 10.67 per order *Notes: FTE = full-time employee Total Cost = actual costs, over the six month period Driver Volume = actual volume, over the six month period Activity Cost-Driver rate = actual rate over the six-month period, calculated by dividing total cost by driver volume Picking a delivery activity = preparation of the sales-orders to be delivered. It basically involves warehouse employees taking, for a given customer order, the appropriate goods from the storage racks and deliver these goods to the quay. From here, orders are combined and picked up by a truck for delivery to customers. Sales-order processing activity = registration of the sales-orders received by phone, either from the customers di-rectly or from Sanacs sales reps. 1.5 Activity complexity When conducting the activity analysis, the controller (Mike Johnson) and his team indeed noticed that many activities were complex. Cost rates, such as the picking cost per delivery and the drop-off cost per delivery, were in fact too aggregate and did not mean anything to De Creus and his sales reps. The sales reps explained that the cost per sales-order processed and the picking cost per delivery were different for each customer because these costs depend on the specific characteristics of the sales-order and on the specific characteristics of the delivery. The following examples describe how these two typical activities were carried out in practice. 1.5.1 Picking a Delivery Sanacs own transportation fleet carries out order delivery. Here we consider the delivery-preparation process, otherwise referred to as picking, i.e. the preparation in the warehouse of products to be delivered to the customer. The total cost (salaries, depreciation, etc.) of the ware-house department in the first half of 2005 was 177,840. The department employed 7.5 full-time employees (FTE) in the warehouse (each of whom worked 38 hours per week, 52 weeks per year), costing a total of 355,680 per year. The employees work at a practical capacity of 80%. The warehouse employees were 100% dedicated to the picking-a delivery activity. In the ABC model, the activity-cost driver identified with picking a delivery was the number of deliveries. However, Mike Johnson was not sure whether this single cost driver ade-quately reflected resource demands associated with the picking activity. His interview data re-vealed the following complexity of the picking activity: The warehouse employee first prints the picking-list. This task is done for each de-livery to the customer (1 minute). Then, the employee drives a fork lift truck to the appropriate storage rack to pick the first sales-order-line, as described on the picking-list (18 seconds). Looking for the appropriate rack is in fact repeated for each sales-order-line. For boxes and containers, an empty pallet is picked (0 seconds) and the boxes/containers are manually loaded on the pallet of the fork lift truck. Loading units (boxes or containers) on the fork lift truck requires 6 seconds per unit. Full pallets are immediately picked with the fork lift truck. Loading a full pallet with the fork lift truck takes 1.5 minutes per pallet. When all products (boxes, containers) are loaded, the warehouse employee drives the fork lift truck to the quay (2 minutes per round trip). Of course, for full pallets the trip to the quay is needed for each pallet. For box-es or containers, the number of trips to the quay depends on the volume of the boxes. When the pallet (full pallet or pallet with boxes/containers) needs to be wrapped-up in plastic film, the warehouse employee needs to stop at the wrap-around packaging ma-chine (3 minutes per pallet to wrap around a pallet). When all goods are ready on the quay, the employee makes a final check and signs off the picking-list (2 minutes). In other words, the interview results show that picking a delivery was driven not only by the number of deliveries, but also by the number of order-lines, the number of boxes/containers/ pal-lets to load, and whether the pallet needs to be wrapped around or not. 1.5.2 Sales-Order Processing The sales-order processing department, which employs 5 FTEs (each of whom works 38 hours per week, 52 weeks per year), carries out the processing of sales-orders. These employees also work at a practical capacity of 80%. The amount of work to handle a sales-order depends on the situation. One of the interviewees described the activity as follows: We are taking orders by phone, either from customers or from our sales reps. Taking an order from a customer is more time consuming than taking an order from a sales rep, because our sales reps are trained to talk in an efficient type of code language and we dont need to do the commercial talk. In general, you can say that the wel-come talk to a sales rep is very short (1 minute), and to a Dutch-speaking customer it is somewhat longer (3 minutes). But for a French-speaking customer the welcome talk is much longer (10 minutes), because these customers tend to be garrulous. Next, we need to open the customer file in the software program. For a well-known customer (i.e., one who places at least 12 orders per month), we know the customer number by heart. Finding the customer file then only takes 12 seconds; for a non-regular cus-tomer finding the customer file takes about 30 seconds, because we need to look up the customer ID-number. If customer data (e.g., delivery address, invoice address, phone, fax, language code) have changed, these fields need to be updated (36 seconds per field). Then, the sales-order is entered order-line by order-line. When the sales rep calls, we can enter 5 order-lines in one minute. However, when a customer is on the phone, we can enter only 2 order-lines per minute. Sometimes, we need to make an additional calculation with a calculator and this activity depends on the number of order-lines (6 seconds per order-line). For rush orders, we need to call the planning department to see whether the products can be delivered within 24 hours (this phone call takes approximately 10 minutes). Finally, some large customers request a confirmation letter. Preparing such a letter and faxing it to the customer takes 15 minutes per sales-order. The interview results show that the number of sales-orders does not drive the activity alone. The time for processing a sales-order depends on the type of phone call received (from a sales rep, a Dutch-speaking customer, or a French-speaking customer), the type of customer (reg-ular or non-regular), the number of customer fields to change (if any), the number of order-lines ordered by the sales rep, the number of order-lines ordered by a customer, the need to make a manual calculation (yes/no and, if yes, the number of order-lines), and the need to write a confir-mation letter (yes or no). Johnson felt that it would be virtually impossible to include all these activity-cost drivers into the ABC model. However, without such detail the resulting ABC model would not be able to estimate the differential resource demands associated with the processing of different sales-orders. 1.6 Time-driven ABC (TDABC) De Creus asked the consultant to look for an alternative to a traditional ABC system. They discovered the time-driven approach, as described by Kaplan & Anderson in the November 2004 issue of the Harvard Business Review. De Creus had doubts about the next course of action: should Sanac proceed with a traditional ABC system (and what kind of ABC system?) or should it implement a time-driven ABC system, which was totally new to him (as well as to the control-ler, Mike Johnson, and the consultant, Chris Miller)? Requirements Assume the role of a business consultant. The case that you have just read summarizes the expe-rience of your client trying to grapple with the issue of cost-system design. Your task is to apply a time-driven ABC model to the firm. As you craft your note, please consider the following: 1. Sanacs business environment (e.g., in terms of markets, products and customers, profitability, and strategy). 2. Why it would be hard for Sanac to implement a traditional ABC system? 3. Build a time equation for each of the following two activities: picking a delivery and pro-cessing a sales-order. For the former, the time equation should estimate the picking time per delivery, while for the latter the time equation should model the processing time per sales or-der. You can find information on the time drivers (and their time estimates) in the interview data in the case. 4. Use your TDABC model to estimate the picking cost for each of the following delivery situa-tions: a. Delivery A (related to sales-order 1): 5 order-lines of 1 box for each order-line, 1 trip to the quay, no wrap-up. b. Delivery B (related to sales-order 2): 1 order-line of 1 full pallet, 1 trip to the quay, no wrap-up. c. Delivery C (related to sales-order 3): 1 order-line of 1 full pallet, 1 trip to the quay, 1 wrap-up. d. Delivery D (related to sales-order 4): 5 order-lines, 3 boxes, 2 full pallets, 3 trips to the quay, customer asks wrap-up of the 2 full pallets and of the third pallet, containing all boxes. e. Delivery E (related to sales-order 2): 5 order-lines, 5 full pallets, 5 trips to the quay, no wrap-up. 5. Use your TDABC model to estimate the sales-order processing cost for each of the following orders: a. Sales-order 5: An order from a sales representative, regular customer, 5 order-lines, no fields to update, no calculations, normal delivery, no confirmation letter. b. Sales-order 6: An order from a Dutch-speaking customer, regular customer, 5 order-lines, no fields to update, no calculations, normal delivery, no confirmation letter. c. Sales-order 7: An order from a Dutch-speaking customer, regular customer, 1 order-lines, no fields to update, no calculations, rush order, no confirmation letter. d. Sales-order 8: An order from a French-speaking customer, non-regular customer, 5 order-lines, 10 fields to update, no calculations, normal delivery, no confirmation let-ter. e. Sales-order 9: An order from a French-speaking customer, regular customer, 5 order-lines, no fields to update, additional calculations needed, rush order, confirmation let-ter is required 6. Comment on the general feasibility of implementing a TDABC like approach in your organi-zation. What specific measurement problems do you anticipate in rolling out such an ap-proach