There are only two really successful platforms in computing today: the internet and Microsoft Windows. By successful, I mean that developers have created software that other developers want (or need) to write software for and computer users want (or need) to use.

Many, many developers try to turn their software into platforms, because it makes their work ubiquitous and it makes them rich like Bill Gates. But few succeed. Some software providers are trying to create software platforms for running your marketing operations, integrating and automating the most important marketing processes together on a single software application.

The lure of such a platform is obvious: Eliminate the complexity and inefficiency of many different—and often redundant—software applications and integrate data in one place to get a single view of all the different things you are trying to do in marketing. There are two primary benefits for doing this, and their relative importance is often determined by the overall standing of marketing in the organization:

1. Improving the productivity and effectiveness of marketing

2. Demonstrating the ROI of marketing activities

But you should beware the platform concept. Windows and the internet are successful because they are infrastructure platforms with limited ambition. They take care of the basic communications but leave the functional capabilities to others. Smart move. There are only so many ways that computing infrastructure needs to communicate. But people are a different story.

And this is the problem with trying to create computing platforms for the ways that people do things. In my experience, it requires a least-common-denominator approach. To create a viable software platform for marketing operations, the developers need to create a software model for the ways that marketers do things that is applicable for the highest possible number of users. And that means inevitable compromises in functionality—most importantly in specificity. At best, these platforms are missing important functionality to serve your marketing processes. At worst, they require that you change your processes simply to be able to use them.

In the end, software platforms that claim to be all-encompassing are a myth. There’s just too much variation out there. And if you do fit exactly to one of these platforms, you should be worried—you probably aren’t very differentiated from your competitors. In 13 years of covering enterprise software platforms for business as a journalist, I’ve learned some basic lessons that can be applied to marketing automation:

1. Don’t do it for integration’s sake. It may seem tempting to throw away all those old software applications and replace them with one, but it’ll never happen. There will always be some important function that falls outside of the capabilities of the platform.

2. Do do it if the platform contains a lot of functionality that you don’t already have. In talking to hundreds of CIOs about their enterprise software projects, the ones who got the most value were the ones who were able to add new business capabilities as a result of the installations—which inevitably run over budget, over schedule, and encounter deep resistance from users who thought the old software worked just fine for their needs. You need to give these people added value for them to suffer the pains of adapting to new software and processes willingly.

3. Have an IT strategy for marketing. Marketing is like most areas of the business when it comes to technology: opportunistic. A specific need arises and you find technology to fill it. That may lead to choosing the best software for the particular job, but not necessarily for the strategic goals of marketing and IT. For example, maybe the software isn’t scalable to other marketing groups, or maybe it is incompatible with existing software, etc. Of course, such issues are what often leads marketing to go around IT in the first place; IT rejects the best software for the job because of all these concerns. In the end, marketing and IT have to have consistent communication about marketing’s goals with technology. That usually means assigning a liaison on both sides.

So how do you start down the marketing automation path? It’s going to sound like a cliché, but you need to take a process view. Chances are, you don’t have a map of all of marketing’s processes, much less the ones that could benefit from some kind of automation. I’ve tried to create a broad list of processes that could benefit from automation. No provider covers this entire list, but that’s not the issue anyway. The issue is, which of these processes are you doing manually today, and which ones, if automated, would provide the most value—not just to marketing operations, but to the overall strategic goals of your organization?

Here’s my list of the major processes:

• Get a single view of the customer. Collect data from multiple places to improve analysis of individual customers
• Model the behavior of the customer so you can predict which ones are the best to do business with.
• Collect and manage conversations about you online and offline.
• Contact customers when and how they want to be contacted.
• Organize marketing content so that it can be targeted at a specific customer, delivered at the right time and in the right context. Automate the delivery of content that supports different customer interactions—call center, sales call, for example—and different events that occur, such as high number of transactions.
• Improve interactions with customers on your website. Can you make your site respond to the customer’s actions and history on the site?
• Better measure and manage marketing activities

Does this cover all of your major processes? What have I left out?

I have a survey out in the field about marketing automation that will wrap up soon. I will be presenting some of the findings during this online briefing in September.