I don't dispute the article's approach of ongoing promotion, but I think there's another way to approach marketing: design the product (and website) to fit the market.
Make the product's benefit easy to see (for target customer task and needs; and in terms that are easy to understand); easy to try it so as to confirm those claims (download, install, "plug-and-play", "getting started", general usability); and easy to purchase (this one isn't so important for enterprise software, because the person processing the sale usually isn't in the decision making loop - bizarrely, it's sometimes a different company altogether that makes the purchase).
The wonderful thing about this is that you are making a kind of sales machine, instead of doing that work manually. It's automation. Of course, it is a tremendous amount of work to actually discover what the market it, who the customers are, what their concerns are, what tasks you can help with and so on. But once you've found out, that knowledge is an asset that you can incorporate into your systems, so it keeps giving value to you - just like a program.
It's sort of "throwing the product over the wall" to see what happens, applied to the marketing aspect itself (except iterative development is needed for both product and "marketing machine").
Of course, this is all immensely appealing to a scientifically-minded technie type, who wants to avoid actually making sales visits or phone calls... I think my approach can succeed, but it probably can't be as successful as someone who tackles sales as a task in itself.
As one data point, I've personally succeeded with this to some extent, but to fully disclose: I do get a lot more sales when I make a release (which is automatically publicized on many websites). I've tried google adwords, trying out different ads, keywords and so on, but it made no impact whatsoever (spent about $100 trying it out).
Make the product's benefit easy to see (for target customer task and needs; and in terms that are easy to understand); easy to try it so as to confirm those claims (download, install, "plug-and-play", "getting started", general usability); and easy to purchase (this one isn't so important for enterprise software, because the person processing the sale usually isn't in the decision making loop - bizarrely, it's sometimes a different company altogether that makes the purchase).
The wonderful thing about this is that you are making a kind of sales machine, instead of doing that work manually. It's automation. Of course, it is a tremendous amount of work to actually discover what the market it, who the customers are, what their concerns are, what tasks you can help with and so on. But once you've found out, that knowledge is an asset that you can incorporate into your systems, so it keeps giving value to you - just like a program.
It's sort of "throwing the product over the wall" to see what happens, applied to the marketing aspect itself (except iterative development is needed for both product and "marketing machine").
Of course, this is all immensely appealing to a scientifically-minded technie type, who wants to avoid actually making sales visits or phone calls... I think my approach can succeed, but it probably can't be as successful as someone who tackles sales as a task in itself.
As one data point, I've personally succeeded with this to some extent, but to fully disclose: I do get a lot more sales when I make a release (which is automatically publicized on many websites). I've tried google adwords, trying out different ads, keywords and so on, but it made no impact whatsoever (spent about $100 trying it out).