An agent-based model (ABM) of a large consulting firm was built by Christopher Stanford at Imperial College London under the supervision of Dr. Wayne Luk. The following work was completed during his Msc thesis and was developed using the Simudyne SDK. This is a theoretical representation of the underlying properties of a consulting firm and provides the opportunity to run Scenario Analysis simulations of the consulting firm.
\n\n\tDownload Model Files\n\nThis paper investigates the modernization of a corporate practice called scenario analysis. It is widely used across the\nglobe, originating from World War II, helping companies to gain a better understanding of potential future outcomes. In\norder to modernize scenario analysis, this paper will explore the novel approaches; of a hybrid development between\nagent-based modelling (ABM) and scenario analysis, the utilization of bottom-up and industry view approaches and\nintroducing elements of human social behaviour. All these features are integrated in a single scenario analysis ABM that\nwill model the fundamental philosophy behind the daily operations of a large corporation.
\nThis paper addresses more natural evolutions for scenarios, by looking at the firms’ fundamental building blocks, rather\nthan just an objective view based on the firms’ financial statements. This model will focus upon accurate\nrepresentations of the daily interactions of an enterprise, especially improving the explainability and transparency of\ncausality in scenario analysis generated by the models, in order to achieve enduring agreements among executives about\nthe firm’s future. With this novel approach to scenario analysis, the modelling of a firm with the use of an ABM, there\nis the potential to provide more in-depth micro-observations of the future, rather than the conventional scenario\nanalysis methods performed today.
\nThis model aims to investigate the novel approach in the modernization process of utilizing scenario analysis, with the\nuse of Agent Based Modelling (ABM). The general application of scenario analysis is to generate best, base, and\nworst-case scenario analysis of a firms’ financial outcome, based upon potential financial and economic inputs, or\nfuture corporate decisions.
\nFor example, this model will focus on people driven services, such as consulting, where some\nof the key variables are the number of available consultants, more specifically the total availability (count), their\nbillable hour rate, working hours, utilization rate (expected down time), current running contracts, future contracts\nand expected contracts, in order to derive expected revenue.
\nSome consultancies utilise a two-pronged approach for its scenario analysis. A bottom-up analysis and an industry view analysis. The bottom-up approach is people-oriented and focuses\non looking at the total amount of consultants available and calculates the total potential revenue\nthey would be able to generate (also referred to as headcount driven approach). The top-down, or\nindustry view approach, focuses on the current contracts in progress, future contracts that are in the\npipeline, and the potential future contracts that may be won with a certain degree of probability\n(also referred to as demand driven approach).
\nA better representation of this approach is shown in fictional representations of the firm in tables 1 and 2.\nThe first representation is achieved by calculating the expected performance of a consulting firm\nbased upon the total maximum utilization of their employees. This can be seen in the fictitious\nrepresentation of the company below in table 1.
\n\n
Table 2 shows how the Industry View Plan provides a further representation of the scenario analysis\nof the company. Table 2 also shows the estimated expected revenue from each company A, B, C, is\n11.5m, 8.9m, and 33.5m, respectively.
\nLastly, both models are compared, as seen in the bottom right of table 2. This shows the discrepancy\nbetween the potential billable income and the estimated client/industry revenue.
\nThis example shows how the expected revenue from the industry view would be higher than the potential revenue from the total workforce. Consequently, to balance the\nbusiness model in this example, the firm would have to either hire more consultants to meet the\ndemand or would have to shift focus towards more profitable sectors and relationships, to maximise\ntheir potential revenue
\nIn order to achieve a bottom-up representation the ABM will integrate the two approaches\ndescribed above. The ABM will be able to automatically calibrate itself in order to maximize\ncompany growth by optimizing its resources and workforce appropriately. This is further expanded\nupon below.
\nThe base model contains the following agents, each with unique characteristics. Employees have project preferences, and consequently a tolerance to undesirable project\nassignments. If the employee is unhappy with the current work environment, they may leave the\nfirm. This becomes relevant if the market model evolves, and the employees specialized in a specific\neconomic sector, which is shrinking, may be assigned to different department projects. If continued,\nthe employee will become unhappy and resign. This is an important aspect that general scenario\nanalysis would be unable to convey. A description of the agents in the simulation is provided in table 3 below:
\n\n
Having designed the key components for the\nmodel to be built around, it is fundamental to\ncapture the characteristics of the model as\nconcisely as possible, while maintaining a\nrealistic behaviour. The main characteristics\ncan be condensed in five main categories (as\nseen in figure 1):
\n\n
Model Initialization:\nThe model allows for ample customization within its starting\nparameters in order to customize different\nstarting scenarios and calibrate the model to\nthe firms’ current operational structure.\nDuring the initialization process the model will\ngenerate the necessary agents and setup the\nlinks between agents as seen in figure 1.
\nContract acquisition & termination:\nThis is the main operation performed by the home\ncompany, where depending upon the\navailability, will choose to either accept a new\ncontract from a client or not. Consequently, it\nwill also terminate and release the assets\nallocated upon termination.
\nHuman resources (HR) management:\nAn important property of an ABM simulation\ncompared to standard scenario analysis, is the\nfact that the model is not static, having the\nability to grow and contract naturally\nfollowing the demands of the market.\nSubsequently, if there is a sudden increase in\nthe number of contracts, the ABM will\nattempt to recalibrate by hiring additional\nemployees, or vice versa. Additionally, an\nelement of human nature is introduced to the\nmodel, where the employees assigned to a\nproject outside of their scope, may feel\nresentment towards the firm and potentially\nleave. This functionality can be further\nexpanded upon to meet the workforce needs.
\nAccounting:\nThis operation is perfomed by the home company agent, which is capable of keeping\ntrack of the P&L of the firm, and of each division, on a monthly intervall (which can be regulated\ndepending upon the different industires). Providing a powerful and detailed representation of the\ncurrent operations of the company.
\nMarket Environment:\nis responsible for managing the interactions between the home company\nand client companies. This is achieved by carefully modelling and simulating market behaviours. For\nthe developement of the ABM described in this paper, the market can be toggled between two\ndifferent scenarios; in the first, the market follows a simple randomised business cycle with a 2%\naverage annual growth (which can be altered by the user), to portray the responsiveness of the\nmodel to changes in the environment. Additionally, there is the possibility to have an exponential\ncontraction or expansion to the market, utilized to stress-test the market and analyse potential weak\nareas in the home company
\nFigure 2 is a graph\ngenerated by the Simudyne\nSDK and shows the links\nbetween the different\nagents generated, and their\ninteractions with one\nanother:
\n\n
The model is designed from the bottom up, emphasising on the fundamental building blocks of a company, rather than looking at the returns from a purely\naccounting point of view. Additionally, the implementation of an ABM, allows to not only simulate\nthe operational interactions within the firm, but it also provides the opportunity to introduce some\nelements of natural behaviour among its employees, in order to introduce an additional element of\nrealism within the model.
","headings":[{"value":"Abstract from paper","depth":2},{"value":"Introduction","depth":2},{"value":"ABM for Management Consulting","depth":2},{"value":"Agent Definition and Interaction","depth":2},{"value":"ABM Network","depth":2},{"value":"Conclusion","depth":2}],"frontmatter":{"title":"Consultancy Model","toc":true,"experimental":null}},"site":{"siteMetadata":{"title":"Simudyne Docs","latestVersion":"2.6"}}},"pageContext":{"absolutePath":"/home/vsts/work/1/s/content/2.5/docs/models/consultancy_simulator.md","versioned":true,"version":"2.5","kind":"docs","pagePath":"/models/consultancy_simulator","chronology":{"prev":{"name":"Factory Simulator","path":"/models/factory_simulator"},"next":{"name":"Financial Tool Kit","path":"/financial_toolkit"}},"lastUpdated":"2026-04-21T13:56:54.856Z"}}