Cost management for development of new therapeutic entities



The current financial systems in place at Eli Lilly and Company for drug development projects are able to report costs (plan and actual) only at the project level, and not at the activity level. As a result, Project Cost Management processes for drug development projects are not well developed. This situation will improve when the enterprisewide system of SAP is implemented for project and portfolio management. The SAP tool allows for projects to be planned and controlled at the activity-grouping level. The SAP team for Lilly spent several months to define the standard work breakdown structure (WBS) and associated standard network of activities for development of new therapeutic entities. The Chemistry, Manufacturing and Control (CM&C) function tested the feasibility of the proposed WBS by conducting a pilot for nine months within the current project time reporting system. This paper will describe the CM&C cost management pilot conducted in 2000 and the associated learning points in preparation for migration to SAP in late 2001. The WBS for CM&C was confirmed to be appropriate. The SAP implementation team is utilizing the learning points as they prepare for the initial rollout of SAP during 2001.

Activity Based Costing for R&D Projects

Activity-based costing is considered to be a project management best practice. However, historically, this has not been done at Lilly. The reasons for this have included: every project is unique; innovation cannot be planned; the attrition rate is high; the portfolio is constantly changing; and so on. However, Lilly is heading into a new era for project management. For over two years functional area managers at Lilly have been developing and refining capacity models for the work required to develop new therapeutic entities. The annual business planning process has been changed from plan-by-function to plan-by-project.Under the new process, individual teams communicate their project plans to the functions, the functions estimate resources needed to deliver the plans on time, and project costs are then estimated after adding administrative overhead. Through tradeoff discussions and analysis of capacity constraints, some project plans, or portions of the project plans, are not accepted into the portfolio of approved work. However, it is difficult to gather these project cost data due to the limitations of the current cost system. Having recognized this constraint, Lilly is implementing an SAP tool that will provide decision-making cost and capacity data to help manage an integrated drug development portfolio.

Exhibit 1. Cost Reporting Levels

Cost Reporting Levels

Exhibit 2. Compliance With Use of New Project Codes

Compliance With Use of New Project Codes

The CM&C functions spent several months in 1999 defining their collective WBS, as did other drug development functions. A pilot within the current time reporting system was conducted during 2000 in order to determine if the CM&C WBS was sufficiently inclusive, and useable for the scientists.While the pilot was being conducted, further SAP design work occurred, and business processes were further refined. The overall scope, schedule, and budget for SAP were also evaluated, and a new plan of staged implementation was adopted by Lilly in late 2000. The initial phase of SAP implementation will focus on high-level deliverables, with additional detail at the activity level to follow at a later time. Exhibit 1 indicates the different levels of detail offered by today’s costing system, the test pilot, and SAP.

The Pilot

Administrative approval was obtained to conduct the pilot for project “Bluebird” within the current project time reporting system.This included approval for establishment of several new project codes to allow for the desired level of cost collecting detail. The project codes were nested under the parent project code, so that the overall cost add-up for CM&C would be comparable to the current system. Functions outside of CM&C (toxicology, clinical pharmacology, etc.) continued to just use the parent project code, while all of CM&C used only the new project codes. Extensive communications were conducted with the dozens of CM&C personnel associated with the project prior to the start of the pilot. The purpose of the pilot and the definitions for the various WBS codes were explained. The team members were located globally, including two locations in the U.S., a Puerto Rican site, and one European site. Thus, a side benefit of the pilot was that it helped the SAP team evaluate implementation issues with a global R&D community.

The team members had to adjust their project time reporting templates to reflect the new WBS activity codes applicable to their work. For instance, an analytical chemist would select “Bluebird—CM&C—Analytical methods development” and “Bluebird—CM&C—all other analytical support” to replace just “Bluebird” on her or his project time reporting template.An organic chemist would select “Bluebird—CM&C—API Commercial process development” and so forth.Monthly followup communications were also sent to remind team members of the appropriate codes to use.

Compliance With the Pilot

Exhibit 2 shows that 83% of the CM&C personnel accurately moved to use of the new project code in the first month of the pilot. Compliance reached 92% by the third month. There was not a compliance problem once a particular person had the correct project codes on his or her personal project time reporting template. Slippage in compliance occurred as new team members were assigned and were unaware of the pilot. For instance, routine assay support is often shared among many personnel within a large department. In the current project time reporting system at Lilly, it is difficult to ensure use of the correct project codes when different people are brought in to work on a project. If one were to conduct a routine search of the available project codes for the entire Lilly portfolio, one would find the “Bluebird” parent project code that most functions were using.Because the current environment at Lilly is to report time only at the project level, and not any functional detail, new personnel assigned to the project did not necessarily search further for the “Bluebird” detailed CM&C codes that were being used in the pilot. As a result, the conductors of the pilot had to diligently review use of the “Bluebird” project codes by CM&C departments and continuously communicate back to the expanding CM&C “Bluebird” team at multiple sites around the world. Due to this continuous communication, by month 9, compliance had reached 98%.

Exhibit 3. Monthly Hours for CM&C High-Level Deliverables

Monthly Hours for CM&C High-Level Deliverables

Exhibit 4. CM&C High-Level Deliverables

CM&C High-Level Deliverables

Learning Point: Ongoing reinforcement communications are necessary beyond the start-up period for an enterprisewide software system such as SAP. Initial training alone will not be sufficient.

Collection of Actual CM&C Cost Data

One purpose of this pilot project was to initiate a database of actual cost for CM&C high-level deliverables, in order to begin to validate the capacity and business planning models in the various CM&C functional areas. The current project time reporting system is able to provide time spent at the departmental level for each project each month. The current system does not allow for collecting the time against project deliverables. The Lilly CM&C organization is built on a functional structure, e.g., formulation scientists are together in one department.Any cost reports from that department cannot distinguish between the cost for true formulation deliverables and the cost for the support of other deliverables, such as manufacturing, or regulatory documents. However, when project plans are developed, the intent is to plan the resource needs against project deliverables. Since Lilly historically has not collected actual data in this manner, project resource plans have been based on best guesses and not hard data. When conflicts have arisen over resources vs. project demands, this lack of reliable data has limited the ability of the decisionmakers to make intelligent capacity tradeoff decisions.This pilot was viewed as an early step to begin building some actual cost data for high-level CM&C deliverables to assist with future project planning and capacity model development.Nine months of data for the eight high level CM&C deliverables used in the pilot are shown in Exhibit 3. Exhibit 4 explains the different high-level deliverable categories that were used in the pilot. The results, particularly the spikes, make intuitive sense. For instance, a major pilot plant campaign was conducted in month 1 to 3 (deliverable E), a technology transfer to a pharmaceutical manufacturing site occurred in month 2 (deliverables B and F), and many clinical trial package orders were filled in the later months of the pilot (deliverable F).

Usefulness of the Deliverables Categories

A second goal of the pilot project was to test with the development team the usefulness of these Deliverables categories in the future WBS structure of SAP. One question that was continually asked of the scientists, engineers, technicians, and operators was,“Does a place exist in the structure of this pilot to accurately capture the work that you are doing?” The answer to this question was:Yes, all the personnel associated with project “Bluebird” were able to find a category to use for charging their time. In some cases, further explanations were needed to accurately identify which high-level deliverable should be used. A catalog was started to collect specific examples of activities that belong under each deliverable, and this catalog is being expanded for purposes of SAP implementation training. For example, Regulatory includes the following:

Proceedings of the Project Management Institute Annual Seminars & Symposium
November 1–10, 2001 • Nashville, Tenn., USA