Several months ago, MonkeyPi blogged about the latest version of RoboHelp (and what he didn't like about it). What I didn't realize was that the discussion is still going on. (I really have to start paying more attention to the comments!)

Scott Abel, the Content Wrangler, wrote a blog entry earlier this week that referenced MonkeyPi's RoboHelp post (actually, he left a comment at MonkeyPi's blog, and then reproduced it on his blog...and no, there's nothing wrong with that). Scott says "..software purchasing decisions ought to be done with a strategy in mind and based on real, actionable data about your environment, your authors, your customers, and your organizational needs and goals."

I agree with him...to a point. No one should choose software based on marketing. (I saw a post to an email list once that said, "I'm buying X because it's [insert marketing blurb here]". Um...shouldn't you make sure it does what you need first?) You need to take several factors into account before sending out that purchase order.

But then he says, "And, the decisions should be metrics-based (quantifiable)...Features are nice, but they have little to do with the true cost of adopting software."

And that's where I disagree. Features are more than nice, and they definitely affect the cost of adopting software. (And wouldn't features be considered part of that "real, actionable data"?)

If your customers need a specific output, the software you choose must produce that output. If your users currently use one application for development, ROI will go down if the software you choose uses a different application, because now you've got added costs for training. If you've made a business case for XML, then you need an application that supports XML.

For example, let's say that you do all the analysis, and MadCap Software's Flare has the best ROI for your team. So you purchase it...and then realize that you can't publish WinHelp, which is one of your deliverables. (I know that this isn't the best example, given that Microsoft has deprecated WinHelp, but the point is still valid.)

Or maybe RoboHelp has the best ROI, but the writers on your team have updated to Office 2007. RoboHelp 7 (currently in beta) will support Office 2007, but earlier versions do not.

I'm not saying to ignore metrics, or to forget about ROI. What I'm saying is that you need to do that analysis on the short list of tools that include the features that are important to you.