5 insights revealed from analyzing 25,537 B2B sales conversations with AI
All to find out what actually works on sales calls
My heart sank.
I had just started my new job at Gong.io, and I was immediately sent to meet on-site with a big potential customer to try to seal the deal (or at least progress it to a next logical step).
My new team was counting on me to get this one closed. It would mean a lot for us.
Throughout the meeting, my level of confidence oscillated up and down as the buying panel seemed to dip in and out of interest.
These guys were tough to read.
And then they said the dreaded phrase…
I asked them their timeline for purchase, and they responded:
Good question. This technology looks incredibly interesting and useful. We still need to figure out who can give the final go ahead on this interally. We’ll be back with you next week.
I know what you’re thinking… Isn’t that a good sign?
Doesn’t it seem like they’re headed in the right direction?
Wasn’t that a “buying signal”?
Counterintuitive Sales Call Insights
By the end of this article, you’ll understand why that bolded phrase struck me with dread.
But I have more to share than just that.
Here’s what I mean…
For you to understand the research and data below, you have to know something about Gong, and the business we are in.
And that business is B2B sales conversation intelligence.
Broadly speaking, we provide the platform that records & transcribes sales calls, maps them to CRM outcomes, and uses AI to analyze which sales behaviors are leading to the highest win-rates.
Here’s why that context is relevant for the rest of this post…
Data-Driven Sales Conversations?
Among 17 of our customers, we analyzed anonymized sales conversation data in pursuit of answering the question: What works on sales calls (and what doesn’t)?
The data was drawn from a pool of well-known B2B technology companies, typically in the mid-market range in terms of size.
We aimed to identify conversational patterns and trends that were leading to the best sales outcomes: higher win-rates, more revenue, and shorter sales cycles.
How We Did It
First, here’s an outline of the analytical approach we used to surface these trends:
- The call recordings of 25,537 B2B sales calls conducted on platforms like GoToMeeting, join.me, and Webex were analyzed — the average call duration was 43 minutes.
- First, we tied the CRM opportunity outcomes to each corresponding call
- Recordings were then speaker-separated and cleanly transcribed from speech-to-text
- Lastly, we ran Gong’s artificial intelligence engine through the calls and transcripts to categorize key sales behaviors and topics discussed
Here’s what we’ve learned so far…
ONE: The Golden “Talk-to-Listen Ratio”
Most sales reps are talking way more than they think they are. The average rep spends 65–75% of the total call speaking, rather than listening.
The highest yield sales conversations hover around the 43% mark of total call time spent speaking.
And it turns out, increasing the customer’s talk time from 22% to 33% radically boosts sales outcomes.
TWO: Let’s Talk Price
The ideal sales conversation follows a predictable pattern in terms of when, and how often pricing is discussed.
When pricing is discussed before value is established (too early), less than 3 times (not enough), or more than 5 times (too much), the chances of closing the deal dwindle.
The golden number of times pricing should come up in a call seems to be 3–4x.
That’s not to say your reps should strive to make sure pricing comes up that many times.
This is more of a buying signal than it is a sales technique.
If you are wanting to take a sales tip away from this insight, it would be this: establish value before you talk price.
THREE: The Signal vs. Noise of Your Prospect’s Timeline
Think back to my story at the beginning.
I mentioned I was “struck with dread” when the prospect responded “we still need to figure out __________________” in response to my timeline question.
It turns out, there is a negative correlation of closing the deal within your expected forecast when the prospect responds to your timeline question with phrases such as:
“We still need to figure out who else needs to be involved.”
“We need to figure out where budget is coming from.”
“We still need to figure out the best internal use-cases for your product before moving forward.”
“We need to figure out _________________” is the kiss of death (okay, that’s an exaggeration).
Of course, let’s remember that correlation is not the same as causation (we know the difference, right?).
On the other side of this coin, we also discovered a positive timing correlation.
When the prospect responds to your timeline question with the word “probably,” there is a positive correlation with closing the deal within your expected forecast.
Counterintutive, but true.
After all, “probably” is not exactly decisive language.
In hindsight, you could make the argument that the prospect is giving a more cautious answer because of how seriously they are considering the purchase.
FOUR: Sooth Their Fears with Risk-Reversing Language
When sales reps talk about (and actively tout) risk-reversing deal policies, win-rates tend to spike.
These are deal terms such as…
- Opt-out clauses
- Money back guarantees
- Easy cancellation
Don’t hide the fact that customers can “get out” easily if they don’t like your product.
Make it plain as day to them.
FIVE: Coach Salespeople at the Conversation Level
There are few higher leverage activities a sales leader can participate in than coaching their reps in a way that improves their sales conversations.
This is where conversation intelligence technologies like Gong come in.
When using this type of technology to perform call recording review and coaching, sales organizations have seen a sharp spike in their sales team’s ability to get deals closed.
We’ve officially entered a new era: one where the sales profession can have as much analytical discipline as the marketing profession.
Now that we have the first wave of this data published, what are your thoughts?
Were there aspects of this article that surprised you?
Or was your intuition simply validated with hard data?
P.S. If you like enjoyed you just read, will you please “Recommend” this post by pressing the ❤-shaped button below?