The Right Way To Do A Software ROI Analysis

On Monday we had a Foundry Group portfolio company sales summit. We are fortunate in that we’ve got a bunch of amazing sales execs in our portfolio, including several CEOs like Howard Diamond of MobileDay and Matthew Bellows of Yesware who have long histories selling and building sales organizations.

The “enterprise sales software ROI analysis” as a selling tool comes up over and over and over again. And most people blow it, or try to bullshit their way through it, or put together something that is clearly not credible. 

So I asked Matthew how he did it at Yesware. Following is his story. Oh, and if you are a Gmail user, check out Yesware.

After spending nearly 20 years selling startup software and services to big companies, I can safely say I’ve seen thousands of “Return on Investment” (ROI) slides. It’s the go-to slide for every enterprise technology salesperson, illustrated with a 4-8 table row, predictably showing that the service in question will pay back the required investment in 6-12 months. Never more (who can wait?), never less (unbelievable).

And like most startup business plans, ROI slides are almost always fake.

The salesperson or their marketing department has no experience to draw on or data from which to extrapolate. Moreover, there’s no accounting for the time value of money, the customer time required to deploy the service, or the risk of time wasted if the deployment doesn’t go well.

Occasionally, a few of the numbers on an ROI slide are based on a previous deployment of the technology. In the rarest cases, the slide has relevant and reference-able data that a potential customer can apply to their situation.

Because of the problems associated with software ROI analyses, we waited a long time to build one at Yesware. And we still failed the first two times we tried. Along the way, we learned that a decent, defensible and compelling ROI analysis requires two key components:

1. Reputable, reference-able customers: The first time we tried to build an ROI slide at Yesware, we anonymously evaluated the data of 40,000 salespeople across a six-month time frame. We were looking for evidence that the people who were using Yesware more actively were making more money than inactive users. Although we found out some great stuff about email open rates and times, and our ROI results looked great to us internally, when we talked to prospects, they were skeptical. Companies, products and industries are so different. No one felt good about applying a broad survey to their specific situation. Lesson learned: Unless a reasonably well-known company is willing to publicly testify to the specific numbers you are showing, you are skating on ice that’s too thin.

2. Identifiable benefits: The second time we tried to build an ROI slide, we worked with one well-known company, analyzing their email and data. We were blown away by the results – a 40% increase in sales productivity between the active and the inactive Yesware users. It was almost too good to be true.

When we presented the findings to the partner company, they were ecstatic. Not because of our results, but because they just had the best quarter in their company history. They were happy to acknowledge that Yesware had something to do with their success, but a successful product launch also played a big role the 40% increase. Lesson learned: Accounting for your benefits should be easy for both the purchasing manager and the finance evaluator to measure. There shouldn’t be too many variables baked into the results.

We tried again, and this time we got it right.

In our most recent ROI efforts we compiled data from three separate companies to uncover the specific benefits their sales teams have achieved using Yesware. These are all well-known companies that our prospective customers can call to learn more – Acquia,Mimeo, Dyn, and WeddingWire. Each is a leader in building modern sales teams, and has offered to be a reference for Yesware.

With this kind of dataset, a simple survey can reveal incredible results. We discovered that on average, sales teams using Yesware:

  • Grew new business (including upsells) by 25%
  • Improved response rate by 32%
  • Improved overall call connection rate by 32%

There are certainly ways to make our ROI analysis better: We will continue to gather a bigger dataset both in terms of customers and salespeople. We will get data from companies outside the USA. And we will keep trying to better tease apart the various contributing factors to changes in productivity.

But overall, we’ve finally cracked the code on a decent, defensible and compelling Return on Investment analysis. I hope this guide helps you create your own.

Matthew Bellows is co-founder and CEO of Yesware, an email productivity service for salespeople. Follow him on Twitter @mbellows.



  • Few points:
    • opportunity costs are not part of too
    • a simple, few variables roi, calculators are less accurate not more accurate
    • the metric out comes mentioned are sales activity outcomes not business outcomes, also %’s mask scope of effort

    I applaud the effort and continued analyses will yield great insights on value, but you may wish to reconsider how you position your findings to prospects.

    Frank Ramirez on LinkedIn in Seattle from my mobile

    • Matthew Bellows

      Hi Frank,

      * I didn’t mention opportunity costs in this post. What’s on your mind here?

      * In general, I agree that more variables in an ROI calculator give a better representation of the levers available to pull. Is that what you were thinking of?

      * In your mind, what’s the difference between sales outcomes and business outcomes? Since we make software for salespeople, we tend to conflate the two. How are they different?

      • Hi Matthew – I got that from “… ’ no accounting for … the risk of time wasted if the deployment doesn’t go well.” The opportunity costs of “time wasted” do not figure into a traditional ROI analyses.
        I built lots of business models and ROI calculators and virtually all require some speculation and different audiences will question those assumptions. In the end the value is likely less in the specific % and more in the education about a deeper appreciation about the weighting of variables makes the business succeed or fail.
        Your metrics talk more to sales activity not too revenue/profit/ margins. Response and connection rate likely have a high correlation to success but its not the same thing as actual revenue. Likewise ” Growing new business 25%” sounds positive but What does that really mean ? the % masks if this equates to $50,$500K, $5MM, or $5B. Marketers hide behind % all the time. You need to qualify the claims more if you seek to gain trust in what is likely a audience needing more convincing.
        I hope that helps 🙂

        • Matthew Bellows

          Totally agree that opportunity costs don’t go into traditional ROI “calculations”. Part of the reason for omitting them is that they are too hard to qualify. But the larger reason, at least in my experience, is that they add to the cost hurdle that the project has to overcome.

          Your second paragraph, especially the part about “the value is likely less in the specific % and more in the education about a deeper appreciation about the weighting of variables” answers your question about us using percentages…

          In addition, our reference clients didn’t want to share specific numbers here, both because they are confidential and because they will be different for other companies that use Yesware. With a percentage, you can compare previous results to your internal performance.

          Thanks for your thoughtful comment.

          • campbellmacdonald

            Matthew: Great post.
            Re: Opportunity costs: They are hard to qualify, but not addressing them in any way doesn’t make them go away especially as unspoken concern in the prospects mind. I’m guessing there is a good story to tell here baselining a Yesware implementation vs other projects.

            We used to lead with “ease of implementation/no tagging” as a leading benefit, which was a mistake as it’s not a reason for buying. But we still speak to it with prospects further on in the consideration phase. Being able to demonstrate low opportunity costs takes away an objection for trying/buying.

  • I’ve built a TCO calculator over the several years for our managed service offering. The key that I’ve found to getting customer buy-in hasn’t been other customer references but instead their own experience i.e. what’s the value of the different outcomes I’m going to deliver.

  • Very helpful. A thought for you, Brad. Any way, you can consider Interviewing the Colorado and Boulder thought leaders like a Matthew to uncover these Gems. Here’s my thought. The reference book I use to gauge investments, companies, and culture is the ‘Art of the Start’ by Guy Kawasaki….No offense to you and your books but Guy resonates on his stuff. What would be most helpful given the New crowdfunding, the Apple ecosystem from WWDC – HomeKit, HealthKit, and the Life-focused blending of consumer stuff into the Digital Media ecosystems (i.e. GoPro IPO coming) is to have Video snippets that Coach, Guide, and Help people wanting to create and operate things. Everything is EASY once someone shows you. 😉

  • Let’s look at this from a Human Nature perspective. If I was going to take a chance a million years ago and venture out from the safety of my CAVE (Colorado) and look to try something new in the world, I’d want to make sure that other PEOPLE were around and that it appeared safe. Maslow’s Pyramid of Technology Needs….the Darwin evolution thing, says 1) Are others doing it(I’m not here to do a wing and a prayer)?, 2) Is it safe (will I get fired?) , 3) Does it solve an addressable problem (My Problem) – If yes, help me and I’ll help you by being a FAN and supporting the TCO calls, etc….In essence, tell a STORY on how the problem was solved but, more importantly, what GENERAL things can be applied to MY PROBLEM. After all, every customer believes they are SPECIAL and they are since they are Human but Nature and Processes are the same now and in 100 years when the next generation is reading a column…..Boulder people live LONG and PROSPER. 🙂

  • Thanks Brad and Matthew. I always enjoy these calculations; especially when diverse data sets are utilized to support a business claim.

    There’s one more piece that always gets me: time to market and competitive value calculations. Ex: “With 25% growth in new business, we were able to land 3 marquee accounts and establish our firm as the de facto brand in our industry.” What’s the ROI of that in 6-12+ months?

    Or one that I like to use here at Rally, “By improving the productivity of your engineering teams by X%, you’re going to generate a great ROI, but that’s a rounding error compared to the ability to get a great product to market first and box out competition.” Note: this is a risky move, and as the authors state, you better have a strong reference customer and/or industry trends to back that up!

  • Thanks for this post. I’m in the middle of an ROI analysis myself, and this is helpful. The conclusion that the results should be tied as closely to your product as possible is where I’m having the biggest challenge. I have a question for you about it, if you would be so kind.

    The result “Grew new business (including upsells) by 25%” seems like it could easily be attributed to other things going on. Response rate and call connection rate seem more attributable to Yesware, and still could be influenced by other things. So, how do you account for the fact that those three results are more attributable to Yesware to your prospects than the 40% increase in sales productivity that was not as credible to your prospects on your second try?

    • Matthew Bellows

      Hi Josh,

      It’s a great question, and it gets to the heart of why ROI studies (especially when they are self serving) are suspect. Everyone knows that there are many contributing factors in any result. Especially in sales, which is complex, non-linear, and subject to emotion, it’s impossible to _prove_ causation. Nonetheless, a good ROI story will make the argument that the service deployed (in this case Yesware) was a major contributor to the company success.

      Our second attempt to develop an ROI study fell short in just this way. Our results were incredibly positive, but the customer had several other contributing initiatives that obscured the role that Yesware played.

      Our third attempt is different because we have three companies that are willing to publicly attribute these benefits to Yesware. So even though _we_ can’t be 100% sure that we created all of the goodness that helped our customer grow new business by 25%, our customers are convinced enough from what they’ve seen that they are willing to testify on our behalf.

      • I see. That’s your two points – the results have to be identifiable to your product and the reputable reference-able customer has to say they attribute the results to your product. Excellent. Thanks.