Machine Learning Reveals 9 Elements of Deal-Closing Sales
The data science team at Gong.io analyzed over 67,000 sales calls/demos to understand the patterns that close deals. Here is what we found.
By Chris Orlob, Gong.io.
It took us 10 weeks of painstaking analysis to do it.
The data science team and I at Gong.io analyzed 67,149 SaaS sales demos to understand the anatomy of demos that close deals.
Every demo recording included in this study was conducted on screen sharing platforms like Zoom and GoToMeeting, where they were recorded with speaker separation.
The recordings were then transcribed from speech to text and analyzed against their sales outcomes (in the CRM) using our machine learning engine to identify common patterns of successful (and unsuccessful) demos.
Data can’t answer every question about demo effectiveness, and every sales organization is likely to have a different demo process.
But that didn’t stop us from finding common patterns that span almost all SaaS demos that eventually lead to a closed deal.
This Is How Long Effective Sales Demos Last
The first thing we discovered is that successful demo calls are 30.5% longer than unsuccessful demo calls (47 minutes versus 36 minutes):
(Note: “Success” was measured by conversion rates to the next stage in the sales process, as well as overall close rates).
Be careful with the conclusion you draw here.
It’s unlikely that effective demos involve reps plowing through every feature their platform offers.
More likely, the duration is due to the demo being so targeted and resonant, it stimulates a robust conversation that makes the call last longer than usual.
Later in this article, I’ll present data that supports this hypothesis. If you think otherwise, let me know in the comments.
For now, here’s the conclusion I’d recommended forming: don’t do long demos for the sake of long demos, but make your demos so crisp and targeted that they naturally prolong the length of the call by stimulating conversation.
Correlation ≠ causation.
This Is the Talk-to-Listen Ratio of Winning Sales Demos
Contrary to discovery calls, where the winning “talk-to-listen ratio” is 46:54, the winning talk-to-listen ratio for a successful demo call is 65:35:
Unfortunately, the average “talk-to-listen ratio” for unsuccessful demo calls is 66:34.
Not enough of a difference to be actionable.
But here’s where it gets interesting.
While the total talk-to-listen ratio during a demo seems to be consistent among successful and unsuccessful demos alike, the interaction patterns that make up the talk-to-listen ratio are wildly different.
While unsuccessful demos still achieve a reasonable talk-to-listen ratio, they tend to have long, uninterrupted pitches that can last up to 106 seconds.
Successful demos are still “rep-heavy” in terms of who’s speaking the most, but they involve shorter “bursts” of pitches that frequently start and stop, allowing the customer to chime in.
We didn’t find a single demo that lead to a closed deal in the analysis that involved more than 76 seconds of uninterrupted pitching.
With unsuccessful demos, long stretches of uninterrupted pitching were a common occurrence.
As a result of this pattern, successful demos have 21% more “speaker switches” per minute, signaling there is much more back-and-forth dialogue taking place than outright pitching.
deliver your pitch in short sprints, not long marathons.
These Are the Four “Acts” of Winning Demos
What surfaced during the latter part of our analysis (which involved some qualitativeanalysis alongside just the quantitative) were four “acts” or “phases” successful demos tend to consistently follow.
Act I: The “Contextual Overview”
Demo calls that lead to closed deals set the context before jumping into the product demonstration.
The topics that arose most frequently during the first 10 minutes of successful demo calls were context-building topics: industry problems, narratives, and overviews.
After the initial small talk and greetings of the web conference, account executives who conduct compelling demos prime the prospect with a “contextual overview,” usually lasting less than two minutes.
They give the prospect the needed knowledge and context for the demo itself to:
A) make sense, and
The best way to do this is by telling the story of the problem you solve, or the narrative that caused your company to come into existence.
Zuora does a great job of this when they explain “the subscription economy” before explaining the details of their product.
Demonstrating a product and its benefits without the required context in place will give the impression that you have a solution in search of a problem.
Act II: The Upside Down Demo
The research revealed that successful demos are most commonly conducted in an “upside down pyramid” manner.
They start with the conclusion, rather than end with it.
Instead of “building up” to the most valuable part of your demo (“saving the best for last”), winning demos begin with the most valuable part.
Peter Cohan, in his book Great Demo!, refers to this as “doing the last thing, first.”
The data agrees with Peter.
In the analysis, a common pattern was that successful demos started by presenting the use case that was talked about the most by the customer during the discovery phase of the sales cycle.
In other words, if a customer spent 7 minutes talking about “coaching,” 4 minutes talking about “onboarding,” and 5 minutes talking about “pipeline management” during the prior discovery call, then the demo would begin by immediately addressing the “coaching” use case, instead of saving that for the end (which is all-too common in SaaS demos).