EP223: ChatGPT: Your New Data Analyst BFF Uncovers Surprising Sales Insights
In this solo episode of Market Dominance Guys, Chris Beall explores the potential of AI-powered data analysis using ChatGPT. Chris demonstrates how this cutting-edge technology can uncover valuable insights from complex sales data in a matter of minutes, a process that would typically take a human analyst days or even weeks. By utilizing ChatGPT’s Data Analyst feature and uploading data from ConnectAndSell, he’s able to quickly examine the correlations between sales reps’ skills and key business outcomes, Chris showcases how AI can help identify the most critical factors influencing pipeline generation and financial success. This eye-opening episode is a must-listen for CEOs, CROs, and CSOs looking to leverage the power of AI to make data-driven decisions and optimize their sales strategies. Join Chris as he delves into the future of sales analytics, revealing surprising findings that could revolutionize your approach to sales training and coaching. Join us for this episode, “Your New Data Analyst BFF Uncovers Surprising Sales Insights.
Key points and timestamps from the episode:
(00:01:37) Chris Beall introduces the topic of using ChatGPT to analyze sales data and uncover insights.
(00:02:39) Chris shares his experience working with ChatGPT to write a book summarizing Market Dominance Guys podcasts in just two days.
(00:04:20) Chris discusses using ChatGPT for a pricing exercise at ConnectAndSell, balancing customer ROI and company profit.
(00:04:55) Chris explains his plan to use ChatGPT to analyze the correlation between reps’ skills and business outcomes.
(00:07:37) Chris walks through the process of uploading data to ChatGPT and having the AI analyze the columns and data structure.
(00:08:34) ChatGPT identifies key columns relevant to the analysis, including activity metrics, conversion rates, and skill scores.
(00:19:49) The correlation matrix reveals that asking for the meeting has the strongest correlation with positive outcomes, while professionalism has a surprisingly low correlation.
Keep listening for the analysis results:
The regression analysis shows that 24.1% of the variability in direct amounts is explained by the combined skill metrics.
The breakthrough score has a significant negative coefficient, suggesting that higher breakthrough scores might be associated with lower direct amounts.
Conclusion:
Chris emphasizes the value of using AI for rapid, unbiased data analysis to gain insights and make data-driven decisions in sales.
FULL EPISODE TRANSCRIPT Below:
[00:00:00] Susan Finch: Welcome to another session with the Market Dominance Guys, a program exploring all the high stakes speed bumps and off ramps of driving to the top of your market with our host Chris Beall from ConnectAndSell and Corey Frank from Branch49.[00:00:20] [00:00:23] Susan Finch: In this solo episode of Market Dominance Guys, Chris Beall explores the potential of AI powered data analysis using ChatGPT. Chris demonstrates how this cutting edge technology can uncover valuable insights from complex sales data in a matter of minutes, a process that [00:00:40] would normally take a human analyst days or even weeks.
[00:00:43] Susan Finch: By utilizing ChatGPT’s Data Analyst feature and uploading data from ConnectAndSell, he’s able to quickly examine the correlations between sales rep skills and key business outcomes. Chris showcases how AI can help identify the most critical factors influencing [00:01:00] pipeline generation. and Financial Success.
[00:01:02] Susan Finch: This eye opening episode is a must listen for CEOs, CROs, and CSOs looking to leverage the power of AI to make data driven decisions and optimize their sales strategies. Join Chris as he delves into the future of sales analytics by uploading his own data and [00:01:20] feeding in complex prompts. This reveals the surprising findings that could revolutionize your approach to sales training and coaching.
[00:01:27] Susan Finch: Join us for this episode, your new data analyst BFF uncovers surprising sales insights.
[00:01:38] Chris Beall: Hey everybody, Chris Beall [00:01:40] here with another episode of Market Dominant Skies. My co host Corey Frank is actually in Armenia right now. And while technically I’m sure it’s possible for us to co-host Market Dominant Skies Armenia and Southern Arizona at the same time. Well, let’s let [00:02:00] him be. It’s getting pretty late there.
[00:02:01] Chris Beall: It’s about 2. 35 in the afternoon here in Arizona. Intuitively, I know that’s got to be very, very late in Armenia. So I want to try something today and let’s see if it works. It could be a little bit tricky, because it could require following a fair amount of verbal input from [00:02:20] me, I guess output, input to you, for something that’s a little bit complicated.
[00:02:24] Chris Beall: But here’s what I want to look at. I want to look at an example, maybe inspirational, maybe more step by step, you can use it in some way or another. to an example of how to use ChatGPT in [00:02:40] particularly to analyze some information that you might have and point the way forward, maybe give you some surprises.
[00:02:48] Chris Beall: So I’ve been doing a fair amount of work with ChatGPT since it came out, at least since I was introduced to it. By a couple of smart young people in the backseat of a car coming back from Whistler, [00:03:00] and they were using it to do something that was kind of surprising to me, but made a lot of sense. Once I understood it better, that was to write a yearly family letter and the results were great, but it was the process that really intrigued me.
[00:03:14] Chris Beall: So I decided to learn a little more about it, ended up working with ChatGPT, so to [00:03:20] speak, to write a book that summarized a bunch of market dominance guys podcasts, and I wanted to see, could you do that in a couple of days? The book wasn’t great. It never got any human editing, but it was usable. You could read it and you could get something out of it.
[00:03:38] Chris Beall: And two days is pretty [00:03:40] short for a book. Half a year, a year, a year and a half, two years is pretty common. So I thought that’s really interesting. And it was very straightforward. As I proceeded through 2023, more and more taking time with ChatGPT to see what I could understand, [00:04:00] uh, building some scoring capabilities for reps skills on cold calling as part of that process.
[00:04:07] Chris Beall: I thought, well, let’s, let’s just see how far we can go with this. And at one point I did a pretty. Substantial pricing exercise for the products here at ConnectAndSell, trying to balance the customer’s [00:04:20] return on investment with ConnectAndSell’s own profit on any given deal, at any given price point, given structure.
[00:04:28] Chris Beall: And what surprised me was that it was a better analyst than any human analyst I’d ever worked with, part of which was, it was just comfortable. There’s something about asking a bot to do [00:04:40] something for you that has both an impersonal quality, but kind of gives you a feeling of, I’m in control of this.
[00:04:46] Chris Beall: There aren’t going to be a lot of, yeah, but I’m not going to deal with a whole bunch of bias. It’s going to make mistakes. I’m going to have to pay attention, but not a bad experience at all. In fact, a very, very good experience. So I thought, [00:05:00] Hey, there’s a lot of things we all believe that might be subject to analysis.
[00:05:05] Chris Beall: And one of them is that RepSkill in conversations is correlated with outcomes, business outcomes. So I thought I’d take a skills analysis done by a GPT 4 model using a skills [00:05:20] rubric itself written by a. GPT 4 model that was provided originally with an input of about one hour of a detailed description by an expert of exactly what needs to happen in a cold call, second by second by second, and what skills need to be demonstrated.
[00:05:39] Chris Beall: So the [00:05:40] rubric was good. It’s kind of long. It describes for each one of the skills that it invented, by the way. It describes what happened at sort of a zero to one point level, you know, two to three and so forth. And it was a 10 point scale. And so just ran that against a whole bunch of data that we [00:06:00] happen to have about our own reps.
[00:06:02] Chris Beall: And pretty much just took the transcripts of all their conversations they had for a few weeks, put them together in, uh, in a form that the GPT. system or the GPT model could digest, apply these skills rubrics to it [00:06:20] and basically score all the reps one week at a time on these different skills. So that was something that was fairly straightforward to do, but it required a lot of, I’ll call it kind of mechanics, you know, grabbing a hold of the recordings and turning them into transcripts and making pipelines of [00:06:40] data that went to processes that did X, Y, and Z.
[00:06:43] Chris Beall: But once I had it, I thought, you know, I wonder, we kind of assume that these skills mean something with regard to outcomes. Why don’t I take a look? So here’s an analysis that was done by the ChatGPT [00:07:00] data analyst, which is one of the kinds of flavors of ChatGPT that you can get nowadays. And I’ll actually read through it and tell you how long it took.
[00:07:11] Chris Beall: So this entire process. It took about 12 minutes end to end once I had all the skills in an [00:07:20] Excel file. And here is my initial prompt. I said, let’s analyze a collection of weekly results of the ConnectAndSell sales team’s cold calls. I will upload an Excel file and walk you through what the columns mean.
[00:07:33] Chris Beall: And we can then conduct a regression analysis to understand the impact of each rep’s skills on [00:07:40] three business outcomes. Meeting set per prospecting hour, meeting set per week, and amount of new sales pipeline influenced directly or indirectly by those reps. Now one thing I like about ChatGPT, I don’t know, I feel a little foolish liking this, but, ChatGPT.
[00:07:55] Chris Beall: com Our emotions count is it comes back [00:08:00] usually with some encouragement. So it says, that’s, that sounds like a great plan. Please go ahead and upload the Excel file. And we can start by reviewing the columns and the data. Once I understand the data set, we can proceed with the regression analysis to examine the effects of the rep skills on the specified business outcomes.
[00:08:18] Chris Beall: So I upload away. [00:08:20] And the data analyst comes back and says, thank you for uploading the file. By the way, no analyst has ever thanked me for anything before. So this is great. Let’s start by loading the data so I can understand its structure and contents. I’ll check the first few rows and provide a summary of the columns.
[00:08:35] Chris Beall: The dataset contains a detailed breakdown of weekly results from sales [00:08:40] representatives activities. Here are some of the key columns that might be relevant for our analysis. And then it lists the columns. The rep name. The week start date and week end date, the dials, connects and conversations, the number of meetings, the total session time, total talk time, total wait time, total wrap time.[00:09:00] [00:09:00] Chris Beall: Those are relevant in ConnectAndSell LAN. Session time is how long we’re using the system to call people. Talk time is obviously talk time. Wait time is the time between when you push a button and get a conversation. And wrap time is the time you spend wrapping up, taking notes, And setting next steps and [00:09:20] formulating your teleprompter for your next conversation.
[00:09:23] Chris Beall: So then it says conversion percentage, conversion rates from dials to meetings and from conversations to meetings. Scores. There was a greeting score, an opener, breakthrough, a meeting score, an objections score, [00:09:40] and a score for assertiveness, professionalism. And there was also an overall score, which was.
[00:09:45] Chris Beall: Actually weighted based on the number of conversations that got the subscores. So then it says there are also more complex fields like direct amount by stage and positive amount by stage, which seem to indicate financial [00:10:00] metrics related to sales pipeline stages. I’ll stop reading here and just say that was pretty good.
[00:10:07] Chris Beall: It got everything right, including the financial metrics. And I’ll explain these financial metrics here in just a moment. Since we’re interested in regression analysis, it goes on to understand the impact of rep [00:10:20] skills on business outcomes. We should identify which specific metrics and scores will be most relevant to use as independent variables, predictors, and dependent variables, outcomes.
[00:10:31] Chris Beall: Please guide me on how you would like to proceed. Specifically, which outcomes and variables we should focus on for the analysis. Now, that was a [00:10:40] fair amount of output for a fairly small amount of input, and it sort of nailed everything. These financial metrics, by the way, are from an attribution report that we have in ConnectAndSell that looks at opportunities by stage, and opportunities [00:11:00] by how they were influenced.
[00:11:01] Chris Beall: So some opportunities came about after a meeting was scheduled. Some after a positive conversation, that was the positive amount by stage, and some after any old conversation, and that’s called possible amount by stage. So all of this data is in the spreadsheet. It’s a very, very wide [00:11:20] spreadsheet, and it has a lot more columns in it that talk about things like.
[00:11:24] Chris Beall: Which objections were the most common? How were those objections handled? And also, what was the rationale for each score? But I wanted to stick to the numbers here. So now I’m about, what, maybe 2 3 minutes into this entire process. I get this [00:11:40] set of analyses and guesses as to what the columns are. And they look pretty good.
[00:11:45] Chris Beall: So I said, well, let’s take a look at the columns with Amount in their name. And it said the columns in the dataset related to Amount are DirectAmount, Direct financial outcome associated with activities. This means meetings, by the way. Positive [00:12:00] amount. These are associated with positive activities. Now, GPT, the ChatGPT got that wrong.
[00:12:08] Chris Beall: It said positive financial outcome. It actually was intended to mean is financial outcome from positive conversations or preceded by positive conversations. But [00:12:20] is good enough for this purpose. I didn’t need to correct it. And then there’s a possible amount, which is the same thing, but it’s possible attribution, possible influence.
[00:12:29] Chris Beall: And then it went on to say these amounts are also broken down by sales pipeline stage. So we could look at every stage and look at, say, the closed one only, or look at [00:12:40] contract through the closed one, or whatever it happens to be. Then it goes on to say these stage specific columns likely provide a more detailed breakdown of financial outcomes according to different stages in the sales process.
[00:12:51] Chris Beall: Given this information. It says, by the way, this is the ChatGPT analyst. Given this information, we could analyze how [00:13:00] these financial metrics are influenced by various factors, such as the number of dials, connects, conversations, and different score metrics of the sales reps. Would you like to explore any specific relationships between these financial outcomes and other metrics in the data set?
[00:13:16] Chris Beall: And I said, we will focus on [00:13:20] direct amount. As the financial outcome of interest. So that means a meeting has to be set. Well, it’s really not shocking that we’re gonna find a high correlation of direct amount to guess what meetings that are set. It won’t be a hundred percent. ’cause not every meeting set is actually a meeting attended, and nor is [00:13:40] an opportunity created for every meeting set, although in our way of doing business at ConnectAndSell.
[00:13:45] Chris Beall: We tend to create an opportunity as soon as a meeting is set, because that way we can track it more clearly. So the data analyst says, great. I love that, by the way. We’ll focus on analyzing the [00:14:00] direct amount as the primary financial outcome to see how it’s influenced by various factors. We can investigate its relationships with 1.
[00:14:07] Chris Beall: Reps performance scores, 2. Activity metrics, such as dials, connects, conversations, and meetings. Thanks. Given the potential complexity and size of your data set, let’s start by [00:14:20] exploring the correlations between direct amount and these various metrics. This will help us determine which factors might have a more significant impact on financial outcomes and are worth including in a more detailed regression analysis.
[00:14:34] Chris Beall: Shall we proceed with this correlation analysis? I love that it asks. If there is [00:14:40] any specific subset of variables or additional factors you’re particularly interested in, let me know. And I said, this is the easiest of the prompts. Yes. That means I’d like to proceed. It immediately outputs a correlation matrix between direct amounts and all the predictors.
[00:14:59] Chris Beall: Of [00:15:00] course, every predictor maps to itself perfectly, so that correlation is one. And then there are correlations of Meetings, as an example, with outcomes and that correlation is very strong with the outcome of direct amount. Meetings correlated at a [00:15:20] 77 level. That’s a high correlation and with conversations at a 59 level.
[00:15:28] Chris Beall: That’s not too surprising. It’s really hard to set a meeting in ConnectAndSell. Unless you have a conversation first. So it doesn’t say that all you really have to do is just [00:15:40] have tons of conversations, but it does say that conversations are highly correlated with meetings and meetings are highly correlated with direct amounts.
[00:15:48] Chris Beall: Therefore, conversations are correlated with. DirectAmount. That’s just kind of the way it goes. That correlation of conversations to direct amounts is 56. It’s almost the same as [00:16:00] the correlation to meetings. So, it gave me this beautiful matrix. So, down one side it says DirectAmount. That’s the answer, so to speak, what we’re correlating to, that’s the financial outcome.
[00:16:11] Chris Beall: And then Dials, Connects, Conversations, Meetings, and then the Scores. The Greeting Score, Opener, Breakthrough. That means how well did they [00:16:20] execute the breakthrough portion of the conversation. The part that goes something like, I believe we’ve discovered a breakthrough. That completely eliminates the need to To do a bunch of manual analysis in order to have a detailed and comprehensive set of insights based on complex data.
[00:16:38] Chris Beall: I just made that up right now, [00:16:40] but that’s an example. And then meeting, which is asking for the meeting and re asking for the meeting and re asking for the meeting. Objections, objection handling, assertiveness, and professionalism, and then the overall score. So this map is red. I’m not showing it to you because we don’t do that here on Market Dominance, guys, [00:17:00] but could.
[00:17:01] Chris Beall: Red is the 1. 0s. That’s a 100 percent correlation. Most everything’s correlated to itself. And then there’s sort of an orange, that’s some lower level of correlation, kind of a white, which is medium level, and then it goes into the cold blues. All the way to, to a dark blue. So [00:17:20] it shows me the correlation matrix.
[00:17:22] Chris Beall: I can look at it right, there is huge value in making a correlation matrix like this in Excel or some other way. I don’t know if I would’ve done it. It’s just too painful. It’s too much work. This is a huge time saver, and I can look with my eyes at it and let my visual system [00:17:40] call things out that are surprising and I say, huh,
[00:17:44] Susan Finch: we’ll be back in a moment after a quick break.
[00:17:52] Susan Finch: ConnectAndSell. Welcome to the end of dialing as you know it. ConnectAndSell’s patented technology loads your best sales folks up with 8 to [00:18:00] 10 times more live, qualified conversations every day. And when we say qualified, we’re talking about really qualified. Like knowing what kind of cheese they like on their Impossible Whopper kind of qualified.
[00:18:12] Susan Finch: Learn more at ConnectAndSell. com[00:18:20] [00:18:20] Chris Beall: Here’s something that’s surprising. There is a very low correlation 0.05, which is almost the same as zero between how professional the reps are, professionalism and meetings, and there’s also a very low correlation. Which is 03, [00:18:40] even lower, between professionalism and the direct amount of pipeline that’s built through meetings.
[00:18:48] Chris Beall: So I wouldn’t have expected that. I don’t think you would either. One would think that. A rep that is more professional, that comports themselves in a way that we would say is professional, they don’t get rattled, they don’t [00:19:00] have emotional responses that don’t make sense, that that would be correlated with positive outcomes, and it’s not.
[00:19:07] Chris Beall: It’s funny, though, because it’s a negatively correlated with the number of connects and negatively correlated with dials, which kind of makes you think that if you have too many conversations, you might lose a little bit of your [00:19:20] professionalism, something to keep track of out there. Assertiveness was a little bit better.
[00:19:26] Chris Beall: It was a 0. 28 correlation out of 1, so 28 out of 100 with the direct amount. So it’s not nothing, but it’s not strong. Objection handling is [00:19:40] 0. 32, a little bit better than assertiveness. But nothing to write home about. And then BreakthroughScript itself. We prize our BreakthroughScript. We think it’s really fabulous.
[00:19:52] Chris Beall: That adhering to it makes a difference. Well, the correlation with direct amount is 0. 17, 17 out of 100. The [00:20:00] opener also we think is super important, and yet it’s correlated only 0. 12, 12 out of 100. Hmm, that’s pretty low, right? Same thing with the greeting, 0. 15. So the greeting The opener and the breakthrough, which is the first kind of chunk, most of the [00:20:20] bulk of the breakthrough script and all of these reps are using our breakthrough script.
[00:20:23] Chris Beall: And these are all cold calls. They don’t amount to much with regard to booking meetings and therefore getting opportunities on the books that required booking meetings. The strongest correlation of all was the [00:20:40] correlation with, guess what, asking for the meeting. And then kind of insisting that the meeting is a good idea.
[00:20:47] Chris Beall: That correlation is 40 out of 100. So it is approximately 2 point something times how well you do in the opener, or the greeting, or the breakthrough, and [00:21:00] it’s about maybe, you know, something like 33 percent more important than objection handling. And assertiveness is vastly more important, uh, you know, more than 10 times, probably 12, 13 times more important than professionalism.
[00:21:18] Chris Beall: Not that we want people to [00:21:20] be unprofessional. I’m just saying it doesn’t lead to measurable results. You probably shouldn’t be coaching it too hard unless you’re just doing it for your own feelings. So, then, I just explained all that to you, your listener, but ChatGPT Data Analyst explained it all to me. And it [00:21:40] basically said, key observations from the heat map.
[00:21:43] Chris Beall: Meetings have the strongest positive correlation with direct amount, suggesting that as the number of meetings increases, the direct financial outcomes tend to be higher. One could say, well, duh, but it feels kind of bad to say, well, duh, to a well meaning bot, right? So I’m going to say, oh, that’s, that’s good that you [00:22:00] saw that.
[00:22:01] Chris Beall: It actually. it is kind of comforting that this obvious connection was made. Conversations and Connects also show positive correlations, though not as strong as Meetings, indicating their importance in influencing financial outcomes. Now this is important in a very simple way. You actually have [00:22:20] control if you’re running an organization to some degree over how many conversations the team has.
[00:22:25] Chris Beall: Your data could be better than it is right now, and it would take fewer dials to get conversations. You could use technology like ConnectAndSell that gets you more conversations. And you could also look at the reps’ activities and ask, [00:22:40] is too much time being spent wrapping up or pausing or hanging out, petting the cat and drinking coffee?
[00:22:47] Chris Beall: Could that time have been spent? Without abusing the rep in actual conversations. Sometimes reps have fewer conversations than they otherwise could. Jeff Blunt says one more dial. [00:23:00] We always say one more conversation here at ConnectAndSell. Then it says most of the score metrics show positive correlations.
[00:23:08] Chris Beall: Those are the skills with direct amounts, but these are relatively modest. This suggests that while higher skills ratings do relate to better financial outcomes, the strength of these relationships [00:23:20] varies. So then it says, given these insights, a regression analysis focused on meetings, conversations, and perhaps some of the higher correlated scores could provide further understanding of their impact on direct amounts.
[00:23:32] Chris Beall: Would you like me to proceed with a regression analysis based on these findings? Or is there any specific aspect you’d like to explore [00:23:40] further? Now, this is super important. At this point, I could go with the flow and say, yeah, that’s about it. But on reflection, what I’m really interested in is. What about the skills?
[00:23:52] Chris Beall: We work a lot in sales on skills. We coach skills. We don’t work as much on data, [00:24:00] even though it could increase the number of conversations easily by getting better data. I mean, say all of your data was for calling was data that told you that these are people who answer the phone. Or their actual propensity to answer the phone.
[00:24:15] Chris Beall: Now we actually have that data in ConnectAndSell and it’s automatically integrated into every [00:24:20] calling session, but it’s still one of those things that management pays less attention to than they do to skills, at least with regard to lip service. I’m not sure how much attention is actually paid to skills.
[00:24:34] Chris Beall: Because I’m not sure that the breakdown of skills is actually done that cleanly. But here’s what I [00:24:40] said to my friend, a ChatGPT data analyst. The correlation with meetings is actually mechanical, in that direct amount is defined as, quote, pipeline that resulted from meetings set in cold calls using ConnectAndSell.
[00:24:55] Chris Beall: This is also true to a lesser degree with regard to conversations and connects. [00:25:00] Even though the score metrics, that’s the skills, have a weaker correlation with direct amount, they’re actually more interesting to me because the impact of skill scores on business outcomes is an unexplored research topic.
[00:25:13] Chris Beall: So let’s try the regression analysis with the score metrics and see what we get. Now, one of the [00:25:20] things when you’re working with generative AI, and I work a lot with ChatGPT and then some other GPT 4 turbo models that we’ve created. is, hmm, it’ll do anything you ask it to do. It’ll give it a shot. This is highly different from most human employees, especially [00:25:40] analysts that you work with.
[00:25:41] Chris Beall: If you say something to an analyst and they don’t really agree that that’s worth studying, you’re not going to get their best. I love the fact that there’s no complaint about rework. or delving in again, or correcting it and giving it another shot by this particular [00:26:00] bot data analyst. I think it’s huge because it lets me be more creative.
[00:26:04] Chris Beall: I can simply ask it to do what I want it to do, look at the results. I didn’t put a lot of mental energy into the mechanics of producing the results because That’s what it did. And I can be more reflective. I can think [00:26:20] more about what I’m seeing and what I’m learning. So, the data analyst here, when I said, hey, let’s do the scores, it says, that makes sense.
[00:26:28] Chris Beall: Ah, comfort to me. Focusing on the score metrics, the skills, to explore their impact on direct amounts will help us understand how rep skill levels influence business outcomes. Since [00:26:40] this aspect hasn’t been deeply explored, it could provide valuable insights. We’ll perform a multiple regression analysis using these score metrics as independent variables and direct amount as the dependent variable.
[00:26:53] Chris Beall: I’ll prepare the regression model and show you the results. Let’s proceed. And then it goes ahead and does the analysis. [00:27:00] Now, if you’re like me, you probably could do a multivariable, multivariate regression analysis. With some effort. Using Excel. I don’t know how you might do it, but you could probably get it done.
[00:27:17] Chris Beall: This wasn’t much effort. I just said, please do it. [00:27:20] And here’s what I get. The results from the regression analysis using the RepSkill scores as predictors for direct amount are as follows. Regression summary. R squared. 0. 241. That means that about 24. 1 percent of the variability in direct amount is explained by the combined score metrics.
[00:27:39] Chris Beall: [00:27:40] Which indicates a moderate level of explanatory power. So this is really important, because even though skills, it turns out, are not highly correlated with outcomes, there are some things we can go after and try to make better. And 24 percent of your pipeline is a lot. So if 24 percent of the [00:28:00] variability in your pipeline, at least in this case produced from meetings, is explained by these score metrics.
[00:28:05] Chris Beall: Well, okay, it might not have been super strong, but it could be worth pursuing. And then there’s an adjusted R squared, adjusted for the number of predictors. And it indicates after accounting for the [00:28:20] number of variables, the model explains about 18 percent of the variance. So less than 24%, still not zero.
[00:28:28] Chris Beall: And think of it this way. If there is potential variance in the pipeline from 1, 000, 000 to 1. 18 million, or [00:28:40] some number like that, the skills could actually influence your ability to get that extra 180, 000 of pipeline. So these are not locks, by the way, because we really don’t have cause and effect here yet.
[00:28:51] Chris Beall: We have correlation, but worth knowing. And then it gives me the individual coefficients, [00:29:00] and the individual coefficients They tell you how statistically predictable these predictors are, that is, how wrong they are. So it says the greeting, opener, objections, assertiveness, and overall score are not statistically significant predictors of direct amount, as their p values are [00:29:20] greater than 0.
[00:29:21] Chris Beall: 05. That is, the p value means probability of the variance just due to Randomness. So, would you get the same thing over and over and over, or would it spread all over the place, so the p value is basically, say, percentage due to chance, eh, kind of high. The score [00:29:40] breakthrough, the breakthrough score has, get this, a significant negative coefficient.
[00:29:45] Chris Beall: with a p value of 043, suggesting that higher breakthrough scores might actually be associated with lower direct amounts, which is an interesting and possibly counterintuitive finding. Now this is the analyst, [00:30:00] the GPT analyst, telling me that this is interesting and possibly counterintuitive. So if you kind of think about that.
[00:30:07] Chris Beall: That’s pretty amazing that a bot that just knows how to spit out next tokens and next, effectively, words, would conclude that this is an [00:30:20] interesting and possibly counterintuitive finding. So I find it counterintuitive also, especially since I’m Kind of one of the people who came up with the breakthrough script.
[00:30:30] Chris Beall: And I would think following the breakthrough well would be a good thing, but maybe not, or maybe it’s only in the case where you’re trying [00:30:40] to follow the breakthrough and you fail that that’s a good thing. So again, we don’t know, but it pointed that out. The analysts pointed that out. The meeting score.
[00:30:49] Chris Beall: This score has a significant positive coefficient with a p value less than 001, it’s a really small chance that it’s due to chance, indicating a strong [00:31:00] positive impact on direct amount. This suggests that skills in securing meetings are crucial for achieving higher financial outcomes. Well, in one way, that’s well dubbed because if you don’t ask for the meeting, you don’t get the meeting.
[00:31:12] Chris Beall: But in another way, it tells you, hey, if you’re going to coach something, go coach that, asking for the meeting. That would, in sales [00:31:20] in general, be considered to be coaching for closing. There’s whole books on this subject, but I think it’s particularly interesting that the mathematical analysis says it’s kind of where it’s at.
[00:31:30] Chris Beall: And then the last one is professionalism, which also shows a negative coefficient. and p value close to the threshold for statistical significance, [00:31:40] suggesting a possible negative impact on direct amount. So, then it gives me an interpretation. We’re almost done. I hope that this was tolerable for you to listen to all this, but I thought it might be interesting because I’m going to go back and just go through the prompts and tell you how long this took.
[00:31:57] Chris Beall: I don’t think I could have done this kind of work if I’d been [00:32:00] left alone for a day, and I wouldn’t have trusted the results. I would have had to go back and check them. More than once. And then also, if I presented them to somebody else, they surely wouldn’t have trusted the results because they go, hey, Chris, these are your results.
[00:32:15] Chris Beall: So, here’s the interpretation. The most influential skills based on this model are [00:32:20] related to handling meetings, that is, asking for the meeting, and breakthrough moments in calls. With meeting skills being particularly critical for financial success. So, interestingly, one was positive and the other was negative.
[00:32:34] Chris Beall: This analysis suggests a selective impact of different skill areas on [00:32:40] financial outcomes, highlighting the importance of specific training and development focuses for sales reps. So now it asks me, would you like to delve deeper into any specific aspect of these results, explore other models, or adjust the analysis in any other way?
[00:32:55] Chris Beall: And I’m going to do a live demo right here, and I’m just going to add a prompt. [00:33:00] Let’s consider The impact on another financial metric, which is possible. Now, I’m testing it here to see if it can figure out that possible means possible amount. It does a great job. It says we will now analyze the impact of the REPS skill scores on another [00:33:20] financial metric, possible amount.
[00:33:21] Chris Beall: Possible amount means some conversation took place, and later,
[00:33:30] Chris Beall: So, what happens here? Interestingly enough, even if the conversation did not [00:33:40] lead to a meeting, you have the same correlations of the skills, the greeting, the opener, the objections, the assertiveness and professionalism to the outcomes, which is pretty much nothing compared to what Except for one thing, [00:34:00] professionalism has a strong negative impact on possible amount.
[00:34:03] Chris Beall: Now think about that. Isn’t that just fascinating? That you can be more professional and maybe you sound a little stiffer. Maybe you’re harder to trust. Anyway, it’s a strong negative impact on the possible amount and on the [00:34:20] amount directly from meetings. So I think that’s quite fascinating, and the breakthrough score is still negative, but it’s not as strong as professionalism, and the meeting score, asking for the meeting, has a positive coefficient on the overall amount, that is, whether it’s from a meeting or not.
[00:34:39] Chris Beall: Now, [00:34:40] we have to have a little bit of warning here. When you’re looking at results from any analysis, whether this kind done by a GPT model or done by hand, By an analyst, you have to be careful of certain things. So this particular analysis has a flaw in it. And the flaw is a kind of selection bias, and that is [00:35:00] the overall metric, which is, Hey, did you get anything in the pipeline at all?
[00:35:06] Chris Beall: Includes the metric from the meetings. That is, from the direct attribution from a meeting leading to the pipeline. So what I should really do is subtract out the, [00:35:20] uh, the pipeline that was generated from meetings and make sure that it’s not being included. And I could go ahead and do that even without reloading the spreadsheet.
[00:35:29] Chris Beall: I could ask the analyst to make up a new metric that subtracts the possible.[00:35:40] [00:35:46] Chris Beall: So again, I’m going to recap the purpose in this whole exercise. The reason I’m doing this is just to show you that there are deep insights, or at least interesting insights that can be gotten much more rapidly than you might think. And I [00:36:00] think in a more compelling way, it’s easier to think about because you didn’t wear yourself out on the mechanics of the analysis, and you didn’t end up with the biases of your data analyst and just using a GPT model, like ChatGPT, and I recommend that the GPT 4 Turbo Model Data Analyst, just because I’m using it here.
[00:36:19] Chris Beall: I’m sure [00:36:20] there are others that are Equally good, and perhaps better, but using that in a simple way by loading up some data and then telling it what you’re trying to do. As I said, let’s analyze a collection of these results and look for a regression analysis to understand the impact of skills. On business outcomes [00:36:40] and just be guided by the response of the GPT model.
[00:36:44] Chris Beall: It comes back and says something, you ask another prompt and away you go. This entire process for me, the first time, I think, took 12 minutes. It would have taken me about a week. I get one pretty picture out of it, which is a correlation matrix, which I [00:37:00] can share with people. I get an analysis that I can share with a little copy and paste.
[00:37:06] Chris Beall: I got some, I can’t do this because the file has gotten stale. So I reloaded the file and did it again. The whole analysis again, I didn’t even have to tell it what to do other than let’s switch over to the possible amount from [00:37:20] direct amount. So, I hope this is somewhat enlightening. The main thing I want to do is, I want to encourage folks to use these large language models as their partner in reasoning through business situations and understanding them without relying [00:37:40] On a human analyst, you might take you a day, two days, five days, or forever to come back with results and may be injecting their own bias, either through selection bias or survivorship bias or something else.
[00:37:54] Chris Beall: And you can avoid a lot of arguments this way, get yourself some insights, and [00:38:00] then they’re packaged up nicely so you can discuss them with others. So a very different episode of Market Dominance, guys, but hey. We believe in data here and unanalyzed data. Well, it’s kind of like an unexamined life. Is it worth living?
[00:38:15] Chris Beall: We don’t know, but unexamined data or unanalyzed [00:38:20] data generally is not worth very much. So for Corey Frank, Enjoying Armenia and Chris Beall, this is Market Dominance, guys.
[00:38:37] Susan Finch: Selling a big idea to a skeptical customer, [00:38:40] investor, or partner is one of the hardest jobs in business. So when it’s time to really go big, you need to use an uncommon methodology to gain attention, frame your thoughts, and employ successful sequencing that is fresh enough to convince others that your ideas will truly change their world.
[00:38:56] Susan Finch: From crafting just the right cold call screenplays, to curating and mapping [00:39:00] the ideal call list for your entire TAM, Branch49’s modern and innovative sales toolbox offers a guiding hand to ambitious organizations in their quest to reach market dominance. Learn more at branch49. com. Never miss an episode.
[00:39:19] Susan Finch: Go to any of your [00:39:20] favorite podcast venues and search for Market Dominance Guys, or go to market dominance guys. com and subscribe.
