Business Application Software Is Not a Product, So Don't Buy It Like One

Treating business applications like products forces buyers to concentrate on selection rather than use while vendors dump resources into costly and irritating sales cycles. Both sides absorb more risk than they should.
Written by Kevin O'Marah, Contributor

Corporate buyers of business applications like supply chain, business intelligence, or product lifecycle management are usually told to find off-the-shelf software products from technology vendors and keep customizations to a minimum. Wall Street likes the idea that a software product, once written, can be copied and sold again at basically zero marginal cost, and thus puts high valuation multiples on software companies’ earnings. The whole system rests on the myth that software is a product like a piece of capital equipment rather than a combination of service and reusable technology components.

By treating business software applications like physical products, buyers consistently make three mistakes:

  • Paying up front to secure deep discounts may make sense financially for products that are hard assets with some resale value, but it does not work for applications that are unique by customer.
  • Ignoring the vendors’ selling expenses is fine when variable costs are mostly materials, but business applications’ variable costs are almost all in the selling, which eventually gets paid for by the customer.
  • Extensive due diligence before purchase can reduce uncertainty for most physical products, but in business applications the only real test is use.
Paying up front leads to the shelfware stalemate

When i2 Technologies reported an average deal size of $1.9M for 4Q00, the industry saw its high watermark for the big deal. A field study conducted by AMR Research at the time found that only about a third of all seat licenses sold were in use. Many companies still have unused software bought years ago, often by previous management, and targeting different problems. Vendors have a tough time selling against the sunk cost of shelfware, users are often stuck with modules that really can’t do the job, and the result is a stalemate.

Selling expenses fall on vendor and buyer alike and are mostly wasted money

The typical enterprise applications field sales person costs about $200K per year. Additional costs of travel and expenses plus technical sales and administrative personnel can bring this number closer to $500K. This expense generally supports three to five potential deals per year, of which 30% close in good times. Much of this expense is mirrored on the buyers’ side with selection teams that may comprise several full-time resources and loads of part-time commitments across user groups. Little, if any, of this spending is directed at making the applications successful in real use, since all focus is on winning the business, whether the fit is good or not.

Risk is magnified, not reduced, by traditional buyers’ due diligence process

The intent of selection due diligence is to reduce the risk of failure and assure a satisfactory ROI. Ironically, the opposite seems to be true. Vendors respond with tweaks to their demonstrations, testimonials from other customers that are already committed, and promises from development that functionality will be delivered in future releases. As vendors struggle to overcome buyers’ objections, they add dependencies and future deliverables that increase the chance of failure while creating a false sense of security for the customer.


Buy business software applications in small, use-tested increments within a broad business strategy. For most corporate buyers, this means more, smaller purchase orders aimed at getting applications into production fast and then learning what works in practice. To be successful, however, this approach needs to be based on a technology platform and process integration strategy jointly owned by IT and a series of business process owners.

Consider the following tactics:

  • Negotiate parameters of the entire relationship, but make incremental commitments for future business. Paying for seat licenses as users go live may seem unrealistic given vendors’ revenue recognition rules, but the principle makes sense and some large customers have successfully done exactly this. The incentive for the vendor is then to focus on getting applications up and running and users trained and happy. This will also reduce uncertainty about future revenue to the vendor, making its high cost of direct sales more digestible.
  • Use pilots. Process automation or decision support tools almost always can be tested in a narrowly scoped deployment. Besides limiting risk, the approach allows for better benchmarking of process improvement and faster organizational learning. This often includes running competing applications simultaneously to pick a winner.
  • Make at least a token purchase early. Your account sales rep has more resources to call upon among their services and development colleagues when you are a customer than when you are a prospect. For many classes of business application, including PLM and supply chain, companies look like long-term turf battles to vendors. Better to have them fighting these battles on the basis of your success and satisfaction rather than trying to do it all in one big, strategic deal.
  • Make decisions about applications and implementation partners simultaneously. Traditional due diligence focuses so heavily on features and functions that more important qualifications like strategy and approach are often poorly understood.
  • Support chosen vendors’ communications with their investors. Wall Street is not going to like the idea that business applications are part service, part product. To soften the blow, it will help if customers provide some visibility on future spending plans. Vendors’ access to capital depends as much on their perceived risk profile as on their earnings potential.
The underlying business model for enterprise software is shifting away from shrink-wrapped applications toward some still undetermined mix of license revenue for reusable software components and specialist services. It will be a revenue stream rather than lump sum approach to winning business. Buyers of such applications should take these shifting dynamics into account when working with sales reps and vendor company executives.

AMR Research originally published this article on 21 August 2003.

Editorial standards