Here, you’ll learn exactly what you need to be able to calculate ROI for your organization’s content management system project. We’ll cover which metrics you should look at, what pitfalls to avoid, and how to determine whether a ROI calculation is even possible in the first place.

The following chapter is excerpted from my book, "Things You Should Know: 25 Lessons I’ve Learned About Selecting Content Technology and Services".

Lesson #3

Sometimes you just can’t estimate ROI on your project

In a perfect world, every organization would know exactly what it would get in return for every dollar it invested in a project. Appropriately, this is called “return on investment,” or ROI.

You probably already know that though, because people have been asking you for it. Your organization might want this number before anyone will agree to the financial outlay.

Sometimes, you can estimate this. Sometimes, you can’t.

And “Estimate” is the only correct word, because no one can know for sure how your project is going to work out, or what effect it’s going to have on the business. It might succeed wildly, or it might fall flat. Anyone who tells you that you’re going to get $X back for putting $Y in is making a lot of assumptions.

You need to measure

The key is measurement. How do you measure the actual uplift in revenue from a website change?

You need three things:

1. Prior measurement, so you have a baseline
2. An identifiable point of conversion¹
3. A method of valuing that conversion

Calculate ROI for an ecommerce project is easy

Let’s consider the easiest possible scenario: e-commerce.

If you’re selling something online, you probably have all of these things:

1. You have some prior measurement in your accounting software. You’ve certainly kept track of every purchase at some level.
2. Your customers perform clear identifiable actions (i.e. they add a product to their cart), culminating when they checkout and actually pay for their purchase.
3. Every checkout is quantitatively measurable. You know how much they spent, what they bought, and how much profit you made.

Under these circumstances, measuring ROI is straightforward. I’m oversimplifying a bit, but you can make a change, then compare the numbers both before and after.

It might be a little more complex if you have multiple changes happening at once, but it’s hard to deny that this is the best-case scenario for determining ROI. The conversion is binary – they either checked out or they didn’t – and you know exactly how much that checkout is worth.

Calculate ROI for a non-ecommerce project is hard

Outside of e-commerce, it’s not that simple. Prior metrics are rare. When organizations want to make a change, they often have to concede that they haven’t been keeping track of much. That lack of reporting and control could be one reason why they need a change in the first place.

Even if your organization does have metrics, the scale of change you’re undergoing might disrupt user patterns to the point where you’re not measuring a “change” so much as you’re creating something entirely new. You’re not moving along a known scale; rather, you’re implementing an entirely new scale. How do you measure the
change in a conversion action that didn’t exist before?

Defining what constitutes a conversion can be difficult. If you have a general corporate website meant to promote your company, do you even have a specific point when a visitor takes a proactive action that provides value?

Best case, you might have a “Make an Appointment” or “Contact Me” webform. When someone submits it, that’s a conversion. They may not have become a paying customer, but they’ve converted from anonymous visitor to sales lead. You can use this to measure the effectiveness of changes, and you can structure your website to maximize this number. It gives you an anchor with which to determine if you’re making things better or worse.

True ROI requires you put some dollar amount on it

Assuming you have this, you can certainly measure your conversion rate, but true ROI requires you put some dollar amount on it. This can be difficult.

• You could track every specific visitor all the way from initial conversion to sale. You would then know when and if they spent any money with you sometime in the future.
• You could track aggregates and extrapolate an individual value. Let’s say for every 100 people who make an appointment, seven of them buy something, and your average sale is $439. Math reveals that 100 people completing that form equals $3,073 in revenue (7 x $439), meaning each completion is “worth” $30.73, more or less.

It’s a far cry from the clarity of e-commerce, but it’s something.

However, this requires comprehensive tracking. You need to make sure you can identify customers who came in through a conversion, and track them all the way to a sale, then somehow compile those statistics automatically. 

When confronted by this, a customer once told me that this would require them to integrate six different systems from three different departments, and there was zero chance it was going to happen.

Multi-channel marketing strategy adds to the complexity

Additionally, it gets complicated when you have a multi-channel marketing strategy. When you’re actively and passively contacting customers through multiple methods, then who’s to say that the website was the specific thing that converted them?

Maybe they saw a display ad while walking through an airport. For all you know, the time they spent filling out the contact form was the very first time they had ever seen the website. And maybe they hated the experience and the website was a drawback that almost prevented a conversion, not an advantage that enabled one.

Things get worse when you don’t have an identifiable conversion. Maybe you don’t have a form that someone can just “fill out” or even if you do, you don’t have a dollar amount associated to that conversion. Maybe your website is just filled with content about your organization or cause, and you’re just trying to plant some idea in a visitor’s head.

The marketing attribution problem

Let’s say you own a construction company that bids on massive government highway projects. No one goes to a website, puts 100 miles of interstate in their shopping cart, then checks out. No one even fills out a contact form. The process of being selected for these projects happens largely offline, mainly through relationships and formal procurement processes.

At best, your website is background marketing. Someone on the contractor selection team might go there just to make sure your company is legitimate. They’ll browse around for a while, review some case studies, and hopefully come away with an impression that your company is one that could get the job done. Months later, while they’re staring down bids and being asked to vote, you just hope that something they saw on your website – or perhaps the overall impression they got – surfaces and influences their decision.²

In this situation, all you’re really “selling” is a memory. There’s no online conversion, and the actual conversion takes place sometime in the future, offline. The best you can hope for is that the memory your digital property leaves behind is strong enough to contribute to a future point of value.

And this is the fundamental problem of marketing attribution, verbalized by the now-classic lament: “I know that half of my marketing budget is wasted, I just don’t know which half.”

Calculate ROI for some CMS projects is impossible

Sometimes, figuring out ROI is just an unsolvable problem. You need good metrics, identifiable conversions, and a way to value those conversions in financial terms. It’s a minority of projects that have all those pieces fall into place.

And at the risk of complete cynicism, understand that some customers have a website only because it would seem weird not to have one. Having a competent website is one price of admission to appear as a legitimate company that your customers want to do business with.

By all means, try to figure out an ROI model and use it to make projections and post-justify an expense, but also be prepared to concede that sometimes, this is not possible. I’ve seen websites contorted far beyond their original intention because someone was demanding an ROI number from a scenario that wasn’t likely to yield it. The resulting mess might have been measurable, but it was likely less effective to the ultimate goal.

Sometimes, your gut feeling on ROI is the best you can do, and you shouldn’t automatically let any misgivings derail your project.

^{1. In marketing, a “conversion” is technically when someone “converts” into a paying customer. More generally, it means when someone takes an action that provides some value to your organization, such as completing a form, making an appointment, purchasing something, etc. What constitutes a conversion is different for every digital property.} 

^{2. I concede that you might use this website for other purposes, like recruiting employees. But if someone is staring down a budget request, they usually want to see revenue upside.}