One of my principal motivations for joining Microsoft four years ago (when the .com boom was still in full swing and Microsoft was a very unfashionable place to be!) was because I wanted to understand what they were doing differently than the other companies I'd worked for or been exposed to.  Lotus made great products when I was there (and may still, for all I know), and yet SmartSuite lost to Office.  CenterLine struggled to a gasping death, never really quite getting it together.  Asymetrix (nee Click2Learn) has still not found its market, and every quarter is a struggle to fend off insolvency.  This general pattern is true at the companies my friends work for and have worked for as well; we all work hard, we build a company, we struggle, and eventually we fail and move on to the next company.  The churn is remarkable; ISV ecosystem is not healthy, and as far as I can tell it has never been healthy.  So I've seen all kinds of ways for ISVs to fail.  I was hoping Microsoft could show me how they could succeed. 

And boy has it been an education working here.  (An MBA ain't got nothin' on the real world, baby!) Here is the first of the three most important “ah-hah!“ moments I've had in the last four years:

• Most software companies totally screw up marketing.  Because of this, your opinion of marketing is probably rather, um, low.  But the failure of marketing is the single most consistent thing I can point to as an answer to the question, “Why did that product fail?“ 

Let me start by defining marketing - I think that will help.  Engineers build what I refer to as the “physical product“ (even though it's only bits - I know - humor me here).  It's the thing that exists in the real world, that you ship to customers, that has functionality.  The role of marketing is to build in the mind of the customer the complementary “mental product“: the clear picture of what the thing is that you want them to buy and how they would use it.  You have to build the mental product before you can get somebody to buy the physical product.  If your customers don't have a clear, effortless picture in their heads of exactly how they'll use what you're providing, they're not going to buy it.

Let's look at a quick example.  Suppose you engineer a pointy metal thing with a handle at one end and tines at the other.  It's the finest stainless steel, it's beautifully engineered, it's got great balance, you're very proud of it, you use it for everything, and you're ready to ship it.  Since you've built it, they will come!  Right?

Okay, so how do you get people excited about it?

It's a multi-purpose tool.  You can use it for all kinds of things!  It can untangle hair.  It can be used as a weapon of self-defense.  It can pry things open.  It can be used as a musical instrument if you bang it on things!  You can use it to pick up food without getting your hands dirty.  It will help you draw a series of roughly parallel lines by providing a guide.  It's even a fashion accessory!

Are you ready to buy it yet?  Yet this is the way most software marketing is done.  In a world in which new markets are constantly being created and new products are constantly being introduced and new features are constantly being added to existing products, the level of information overload is spectacular.  When we turn around and add to the information overload and expect our potential customers to take the time and energy to figure out from the breadcrumbs we drop what a possible scenario is in which they might use our products, we deserve what we get, which in most cases is failure.  And if you're an engineer, you can and should point the finger at your marketing people, but you also need to understand that the engineers often play a role in this fiasco by insisting that every possible use of a product be enumerated in every communication.

Instead, we need to be clear and simple in building the mental product.

It's a fork.  You can use it to scoop up a bunch of little food or stab and cut big food while you're eating.  It's polite to use it; it keeps your fingers from getting messy, and allows you to cut food into smaller bites.  When you're eating little things like peas or rice, you can pick up a whole bunch of them at once with the fork, making it easier and quicker to eat them.  So it makes eating a lot of different kinds of food easier and less messy.

But in order to tell that story, we have to choose.  We cannot tell the story of every possible use of a fork.  We cannot talk about all of the cool features of the fork (“See how the stainless steel gleams in the light!“) in the absence of their direct applicability to something people might do with the fork.  And we can't change our minds with every communication about what story we're going to tell about how a customer should imagine using the fork.

There is a lot more I could say about basic positioning and messaging, but that's it in a nutshell.  The hardest thing, quite frankly, is having to choose.  And Microsoft has mechanisms in place to force marketers to choose their messages, to clarify them, and to communicate them consistently internally and externally.  It's a huge part of why Microsoft's marketing has traditionally been so much more successful than the industry's at large.

I won't go into detail about my other two big ah-hahs, but they are in brief:

• Strong program management makes a huge difference - these are the people who are responsible for specs and schedules, for making sure you're building the right product for your market, and that your timing and feature set are optimal.  This is a different discipline than development, test, or marketing, though it has significant interaction with all of those disciplines.
• Microsoft has a core strength in business model strategy.  The question to ask yourself (in addition to the obvious, “Gee, do the numbers work out at all in even the most optimistic scenario?“) is , “How will the 1000th sale of this product be easier than the first sale, and the 10,000th sale be easier than the 1,000th?“

Most software companies are capable of putting together code that will compile.  Where they fail is in making sure they're building the right product, that their customers have a very clear understanding of how and why they would use that product, and that the way they sell and get paid for the product will actually scale.