It’s hard to predict popularity

By far, the most popular of my free utilities is doc2html, which converts (simple) MS Word documents into relatively clean HTML. It's downloaded around 150 to 200 times per month, and there've been peaks with upwards of 1,000 downloads in a month.

In contrast, my next most-popular program is Count Anything (a word-count program), which gets around 30 downloads per month, with a peak of around 200. Following Count Anything are PowerPoint2Doc (25 dl/month), and WordSpeaker (20 dl/month).

Programming effort

In terms of programming effort, I would rank these applications as follows (highest effort to lowest):

  1. Count Anything
  2. PowerPoint2Doc
  3. WordSpeaker
  4. doc2html

I wrote doc2html over about a week, spending a couple of hours per day. I wrote Count Anything over a period of several months, probably spending about 100 hours total. In fairness, one reason why I was able to write doc2html so quickly was because I reused most of the Word document parsing code from my TransAssist application.

Personal usefulness

In terms of personal usefulness, I'd rank them as follows (most useful to least):

  1. Count Anything
  2. WordSpeaker
  3. PowerPoint2Doc
  4. doc2html

I actually use Count Anything and WordSpeaker just about every day — Count Anything to invoice PowerPoint and Excel translations, and WordSpeaker to proof my translations in MS Word. I used to use PowerPoint2Doc a lot more, in order to get word counts of PowerPoint files, but Count Anything has taken over that role.

I've never used doc2html outside of testing. In fact, while I wrote the other three programs in order to solve actual problems in my work, I wrote doc2html for a software localization workshop I was teaching. I thought it'd be useful to localize an actual program right in the workshop, and that it'd motivate the students if I published the localized version along with their names.

I never thought that doc2html would be popular; I wrote it purely as localization fodder. And I wasn't the only person who thought that; at the workshop, one of the students asked me if anyone would actually want to use such a program (I suspect she was worried about attaching her name to a dud). So it was a little ironic that doc2html turned out to be my most popular free utility ever.


In terms of programming effort and usefulness to me personally, doc2html is at the bottom of my four most popular free utilities. Yet it's by far the most popular of them, at least in terms of number of downloads.

One reason for this is probably market size; I'd guess that a lot more people want to convert MS Word documents into HTML documents than there are people who want to get word counts from strange file formats, or proof-listen their Word documents. On the other hand, there are plenty of free utilities for converting Word documents to HTML, and as far as I know, none that do what my other utilities do (that's why I wrote them).

A lot of the common wisdom you hear about small software projects is to find a niche; huge markets have huge competitors, but if you find a niche in your own area of interest, you can become the solution. I would have applied that wisdom to say that Count Anything should be more popular than something like doc2html. I suppose that while the Word to HTML conversion market is a lot larger than the word count market, it's still small enough to constitute a niche.

I guess my take-away lesson from this is that market size (big but not too big) is a better predictor of application popularity than programming effort or the usefulness that you personally find in the program. On the other hand, since I'm writing these utilities for my own pleasure, I don't see how that could help me decide which utilities to write in the future…

2 comments to It’s hard to predict popularity

  • Ben

    Is it possible you could add some comments or something on the download pages for your utilities for the benefit of people who want to ask questions? These macros are very useful, but I’ve noticed two bugs in FindNextJ and one bug which can crash Word in Wide2Narrow.

    The bugs in FindNextJ are
    * For some reason it always highlights the very last character in the file
    * It doesn’t seem to turn off screen updating

    The bug in Wide2Narrow is because of the algorithm which seems to replace everything in the document, including numbers which aren’t fullwidth ones. If you have the corrections button turned on (変更の履歴) it actually crashes Word. Perhaps it would be good to document that somewhere.

    I haven’t used Word2Html but I know that several people find your Japanese character macros useful so I hope you will go on improving them.

  • Thanks a lot for pointing out these issues, Ben. In the past I’ve simply asked people to email me at software at ginstrom dot com with any bugs they’ve found, but some kind of online comment system would probably be a good idea.

    What would you think of something like a Wiki?

    In the meantime, I’ll get these bugs documented, then try to work on some fixes. Thanks again!

Leave a Reply




You can use these HTML tags

<a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>