November 18, 2003

Predicting Software Success

Some new applications are destined for failure, others for greatness. Here's how to tell before you see a line of code

by Joshua Greenbaum

Continued from Page 1

Replaces Spreadsheets and Whiteboards

This attribute is a corollary to the "keep it simple, stupid" message of automating existing processes. Some of the best ideas for innovative software come from wandering around companies looking for an overabundance of spreadsheets or observing a roomful of people looking at whiteboards and scratching their heads. Once a spreadsheet or whiteboard is used on a regular basis by more than three people, and draws on data from more than two disparate sources, someone should start thinking about coding the processes and data into an application.

The spreadsheet/white board replacement imperative becomes even more acute as the processes span divisions and geographies. A relatively efficient local process that works well enough on a spreadsheet starts to get very inefficient and unautomated as the process is extended up and out. This situation is where that often misused buzzword — collaboration — starts to make sense. If you have multiple offices full of people trying to solve a companywide problem using a whiteboard or a pass-around spreadsheet, just automating the collaboration itself could make a huge difference.

Software that keeps the whiteboard and spreadsheet processes functionally similar, while automating things such as data storage, workflow, business rules, and collaboration, can mean the difference between software you want to own and software that wastes your time and money. My favorite example here is consensus forecasting: If you look at what start-ups such as RubiconSoft Inc. and Steelwedge Inc. are doing in forecasting, you'll find their specialty is tearing down the whiteboards and automating the spreadsheets. The moral here is also simple: Automate the obviously unautomated.

Automates Business Objects

It's interesting how many of the applications categories we currently use are fundamentally about automating the processes that surround the use of key business objects. Financial software focused originally on the general ledger, human resources software automated the employee record, customer relationship management automated the customer record, and so on. Every relatively stable business object or record that sits at the nexus of a set of business processes is an object ripe for automation.

Once you understand that, figuring out from where the next software categories will come is simple: contracts, bills of material, and service-level agreements are among the many classes of business data or object that need a little more automation. Start-ups Encover Inc. and Determine Software Inc. are both focusing on contracts as a key data object: Turns out terms and conditions are key data objects that no one has bothered to automate before. The trick isn't automating the document management side, but automating the tracking of terms and conditions found within the contract. The moral, once again, is a simple one: The more you reuse a given data object, the more you should automate the processes that revolve around it.

Does this checklist always guarantee success? Absolutely not. You will always need to buy applications that do things you've never heard of or seen before, and some applications will match these criteria and still not be worth your time. But as the number of new applications continues to grow in anticipation of an uptick in buying, everyone should at least be aware of what works best. Granted, following these three criteria ends up almost guaranteeing that change happens more slowly, and existing investments are preserved for a longer time.

Rate This Article Rating: 1 (best) 2 3 4 5 (worst) Comments: Optional e-mail address: