Search | Site Map
   

Changes in Business Models      

       



As the world moves to electronic commerce, some of the most important changes are occurring in what businesses do and how they derive value from doing these things, in other words, in what business models they use. In many well-known e-commerce success stories -- such as amazon, eBay, and Dell, for example -- the fundamental innovations were not just using new technologies but using new technologies to enable new business models. An important part of this research project involved studying changes in business models.
We define a business model as a description of the activities that a company performs to generate revenue or other benefits, and the relationships, information, and product flows a company has with its customers, suppliers, and complementors. We propose to study changes in business models in three ways:
developing frameworks for analyzing and classifying business models
classifying empirically the business models of substantial numbers of companies
testing hypotheses about the distribution, performance, and evolution of different business models

Research Sites and Methodology

The goal of this sub-project is to study business models in the US economy and how information technologies influence the relationships an enterprise has with its customers, suppliers, and employees. Our focus is to understand, classify, and measure the current and changing business models of enterprises. This research draws heavily from our previous work on business models and business processes (see Weill and Vitale's book, Place to Space or use link on Publications page), which identifies eight elementary electronic commerce business models, and the Process Handbook, a structured on-line database of business activities and processes).

Over the course of the this grant, this project

  • Developed and refined, through an iterative process, a taxonomy of business models. As shown in Figure 1, the classification scheme, a four-by-four matrix, has two axes: asset rights and asset types. The four types of rights are: (1) ownership (of an asset which has been transformed significantly by the organization whose business model is being classified), (2) ownership (of an asset which has been transformed in a limited fashion), (3) use of an asset, and (4) matching of a buyer and a seller for the asset. The four types of assets are: (1) financial, (2) physical, (3) intangible, and (4) human resource (HR). We believe that all businesses can be naturally classified into one of these categories, and that each category includes more detailed sub-types.

  • Elaborated a set of rules that relied on publicly available firm data for manually classifying the business model of a firm; these rules eventually became the basis for manual classification efforts and the automated classification tool.

  • Collected a variety of financial data for classification efforts and data analysis, including information about a company’s balance sheet, income statement, cash flow and financial ratio data from COMPUSTAT, Wharton School of Business, finance.yahoo.com, edgar online and other readily available sources.

  • Initially classified the business models of the 1000 largest firms in the US economy (also know as the SeeIT 1000), using FY 2000 as the base year. Increased the coverage of the original 1000 firms to cover all publicly available data for fiscal years 1998-2002 (4,876 analyses).

  • Extended the automated analyses to cover all publicly traded firms for 20 years that have Compustat revenue data, which yielded 95,786 automated analyses.


Findings

It appears that:

  • A relatively small set of simple decision rules captured a significant portion of the distinctions needed to classify companies.

  • Four basic business model archetypes can successfully classify all the revenue streams of all the companies analyzed, however, a matrix of 16 models is more complete.

  • Because only a relatively small set of simple decision rules were needed to classify companies, this allowed us to develop, test, and refine an expert system for automated classification.
    • Based upon descriptive information related to revenue segments and SIC codes, rules are applied with decreasing weight:
      • Comparing segment descriptions year by year (same company)
      • Comparing segment descriptions from other companies.

      • Comparing NAICS or SIC codes for similar descriptions from other companies

    • The classification accuracy rate for the model classification is over 90%.

Based on a five-year (1998-2002) analysis of 10,970 of the largest companies in a diverse range of industries, the results are as follows:

  • The distribution by percent of total firm revenue by Basic Business Model Archetype (in FY2002) was as follows:
    • Creators 49.6%

    • Landlords 34.3%

    • Distributors 15.2%

    • Brokers .9%

  • The percent of total revenue in each Basic Business Model Archetype did not appear to vary significantly over this time period.

  • The distribution by percent of total firm revenue by asset type involved in the business model (in FY2002) is as follows:
    • Physical 73.6%

    • HR 9.9%.

    • Financial 13.5%

    • IP 3.0%

  • The dynamic view over the five-year time presented an interesting reduction in the number of business models, from 9,563 to 7,123, suggesting that firms seem to be consolidating in terms of business models.

  • Each company classified in the sample may have revenues from the use of multiple Asset Types. Of the 10,970 firms in the five-year sample, more than 50% produce revenue from only one Asset Type.

  • The most common Business Models (in FY2002) are:
    Business Model Archetype% of Total Revenues
    Manufacturer (a Creator selling Physical Assets it has transformed significantly) 48.9%
    Contractor (a Landlord selling use of Human Assets) 9.9%
    Wholesaler/Retailer (a Distributor selling Physical Assets it has transformed negligibly) 15.2%
    Financial Landlord (a Landlord selling use of Financial Assets) 12.8%
    Physical Landlord (a Landlord selling use of Physical Assets) 9.1%
    Intellectual Landlord (a Landlord selling use of Intangible Assets) 2.9%
    Financial Broker (a Broker selling the matching of a buyer and a seller of Financial Assets) 0.7%

  • Business models matter. In a joint test of the significance of business models on performance, the p values for all six measures of financial performance, except for margin on sales, are significant.

  • Some business models performed better than others on specific measures, although none were superior across all performance measures.

  • Business models—in combination form—matter (p-values on overall business model combinations are significant). And of the 258 coefficients in the table (these are for model combinations with at least ten observations), 49 are significant. Taken together, we interpret this as evidence that some business models combinations are superior to others. Yet no combination is superior on all measures of performance.

  • Business model effects are greater than industry effects––6.3% versus 5.6%—and these effects are robust.

Sub-project Investigators

Thomas Malone
Peter Weill


     

Home | Research | Contact us | Publications | Affiliations | Abstracts | Links | Search
Introduction
Current Projects
Organizational Practices
Business Models
Organizational Design, Structure and Performance
 People
 Alumni
 Directions and Map
Center for Coordination Science
Center for eBusiness
Center for Information Systems Research