I don't know if this will be useful to anyone, but I did well a million years ago, back when msDos was an adolescent, windows a zygote and CPM had missed the boat.
The selling situation was a little different, and the competition was as much a mindset as other suppliers. People resisted change having always done it "this way" where "this way" was things like Kalamazoo cards, Roneo machines, and etc.
But some principles will still apply.
I developed an approach and a set of routines that were really user friendly (for those days) and build a development foundation (like Btrees, etc) and then used them as the core for everything.
I then targeted vertical markets. I learned as much as I could about a market in general, then approached people who were ambitious and/or influential in the that industry, with a concept for a system tailored to the industry.
In return for their advice on the fine details and the use of their influence in the industry, I cut them in on the action one way or another. That was negotiable, depended on circumstances, and ranged from commissions on sales to partnerships in JVs.
The software was bundled with equipment and sold as a package.
When my business was established enough, the JV approach was not as necessary, but it still paid to have people around who might know zero about computers but knew what the marketplace wanted, and the ins and outs and mindsets of the people in that market.
I did well in some markets (capturing up to 70% of one - paydays!) and reasonably well in others.
I also had a bread-and-butter line based around standard inventory, AR, AP, etc.
After a few years I made the best move of all and sold off that business.
I believe the approach still has some validity, and can be used now, albeit with different medium and in a different environment. I intend to find out, and the challenge is interesting and invigorating. (I also want to conquer again, in my own right and not through one of my companies or businesses - before I move into my twilight years.

)
Anyhow, the approach is to find something that is lacking or poorly addressed (the "vertical" market or the "specific" need) and provide something that works and is
very easy to use.
The purists will hate this, but the user is much more interested in (and judgemental about) the bits they can see than the real stuff doing all the work behind the scenes. Only programmmers really care whether the code is 6 or 8 lines long, or takes 10 extra milliseconds after a month of running. So provided the package delivers the goods and does the job
well enough and is simple to use ...
Anyhow I am banking on it. I believe I have identified a few areas where there is potential, and where a momentum can be built starting with one-on-one sales and installations.
What I have to and want to learn is internet marketing, mass packaging, etc, because I have no idea how to sell or market online.