Choosing a business software package such as Microsoft Dynamics, Salesforce, SAP or Oracle is a daunting task, and the process can lead you astray in making the best decision for your organization.

There is no shortcut to finding the right product.  Choosing the best selling or fastest growing software product does not guarantee you will be successful in your company with that product.

Software categories are complex, and many products are geared toward specific industries or to companies of a certain size.  If a small company tried to implement a product such as SAP enterprise resource planning which is successful with large global companies, it could easily find itself overwhelmed and the implementation price could exceed annual company revenue.  On the other hand, a large company with global subsidiaries, complex tax rules, and high number of users and transactions would not likely be successful with QuickBooks.

One of the approaches to choosing software is to create a list of desired features and send solicitations to vendors in order to find the ones that best fit the business need.

My company InfoStrat participates in many product evaluations, and these are the top pitfalls that I have witnessed:

  1. Vendor feature lists are rigged to make that vendor look good. They are created by marketing staff, and no vendor will include features that they lack on a feature checklist.  As a bidder, if I see a solicitation that contains a feature list taken directly from vendor literature, I know that a product has already been selected.
  2. Asking vendors to rate themselves on features results in optimistic ratings.  Vendors know what customers want to see, and it is tempting to answer Yes to everything unless it is clearly not supported.  The more vague the requirement or feature, the easier it is to justify answering Yes. You could mitigate this problem by asking third parties such as resellers to answer as well.  This will not likely result in consistent results, but could point you to features that need further elaboration or examination. 
  3. Features are an arbitrary black hole.  Defining what is a feature is arbitrary, and you can come up with a large number of features for any product.  More features do not necessarily make products better.  I prefer to carry a Swiss army knife with only 5 features because it is lighter and more comfortable in my pocket than the largest model. Software can become bloated as well, to the detriment of usability. Some features are redundant or overlapping, and can lead you astray when you favor the sheer number of supported features.  For instance, for an automobile you could ask whether it contains cup holders (1 feature) or break this into driver cup holder, passenger cup holder, left rear seat, right rear, middle seat and third row cup holders.  Next thing you know, you end up driving a minivan.
  4. Counting features without weighting them distorts the evaluation.  This is closely related to the previous point but surprisingly common.  If you count positive and negative answers to a features list without applying weights to the more important criteria, you are weighting them equally.  This can lead to poor decisions.  Consider the hundreds of criteria you could apply to purchasing a vehicle and ask yourself whether you should equally weight criteria such as safety, affordability and capacity with side mirror defrosting or the interior mood lighting. 
  5. Favoring objective over subjective criteria.  It is tempting to reduce a decision to what seem to be objective criteria, removing subjective criteria or reducing their importance.  The problem is that for business software user adoption is key, and adoption is heavily influenced by criteria that resist being reduced to feature check boxes.   We see evaluation matrices with criteria such as "intuitive" or "seamlessly integrated".  These may be good aspirations, but they are impossible to evaluate objectively.  
If a feature list is arbitrary, rigged, and easily distorted, what is a better way to choose enterprise software? Our most successful customers have used requirements and features to narrow their choices, and then conducted a proof of concept or pilot to allow their users to get hands on experience with the products.  This approach allows you to factor in subjective criteria and determine the fit of software with your organization, from both a technical and a business perspective. 

Cloud software is much easier to experience in a pilot than on premises software because it doesn't require you to configure servers or your network to start using it. Many vendors offer free trials for 30, 60 or 90 days and they will usually extend a trial if you need more time to make a decision.