Loading summary
Transcript54 lines
- [00:00]
A
Today we're going to turn you into a data analyst. That might sound boring, but it's not. AI makes analyzing and understanding data possible in ways that's never before been achievable. And why you have to watch this show is because I have an ex Google data scientist who has decades of data analysis and data science experience. She is going to give you the cheat sheet. She's going to give you all the tips, the tricks and the guidance to actually do this for your own company. This is going to be a game changer. It's going to unlock your career and unlock your growth. Let's get into today's show. What we're going to talk about today is agentic analytics and how you actually use these AI tools to do really powerful and meaningful data analytics. And today we're going to go over Codex. And so maybe before we actually get into the demo, like, give us a little bit of like, why Codex? What's your take on Codex for data analysis and give us the unvarnished opinion.
- [00:57]
B
So Codex and ChatGPT have come a long way. Like for those of you who have been following ChatGPT launches. A couple of years ago, ChatGPT launched something called ChatGPT operator. They literally called a data analyst. And I tried it out and it was horrible. It was not good at all. It was doing a very poor job. So compared to that, I would say like OpenAI has come a long way with respect to data analysis. It is not perfect, but it the hard coding part, like with the data analysis specifically and crunching numbers, I think that is where ChatGPT and Codex has come a long way. For example, the demo that I'm going to show today is I basically have a CSV file which I take and put it into Codex and I give it a problem that I want trying to solve a root cause analysis and you're going to see that it does a fundamental job. Now does that mean I'm going to take that at 4 face value? Definitely no. I'm going to do some validations on it. I'm going to make sure whatever is spitting out is actually correct.
- [01:55]
A
Before we go forward though, Codex can seem like the scary thing to people. Yeah, I think we're obligated to say that if you have a ChatGPT account, then all you have to do is go download Codex and it's a very quick setup on your computer. And so it's not this big scary thing. It's like a couple of minutes to get up and running. Right.
- [02:14]
B
So like just like Google or like watch some YouTube videos. They're like awesome content online that takes you like from very beginner to like somewhat intermediate, where you're comfortable to like just get started and open Codex and like put your analysis file and like run the analysis.
- [02:30]
A
I did a great show with Matt Wolf where he built like a, an AI second brain in Codex and we did something Codex set up there so you can go check that out. So this video is going to assume that you have that part done, right? You got Codex and you have that set up. And then if that's the case, then we can kind of go to the data analysis use case. And so I know that you were talking about that you're going to walk us through a case where you have data in a CSV and so take us from there.
- [02:55]
B
So this is the file that we're going to be working with. This is customer retention data. Basically we have visit id, we have customer id, their signup date and when they're visiting. And. And we also have like their visit date, like what week they visited, what were their total number of visits, Are they new customers? They're returning customer? Basically it has like all customer retention data that you would expect in a customer retention dataset.
- [03:21]
A
Retention is this complex topic, Right. Because it's not just like binary, like, I was a customer, was not a customer. It's, well, I'm still a customer, but I pay you less. I'm still a customer, but I pay you more. I. I'm still a customer, but I bought this other product. Or I'm not a customer of this one product, but I still retain this other product. It's a very like, multifaceted, complex topic, which I think it makes it really good for the kind of demo and example you're walking through.
- [03:44]
B
Exactly. Like your best customer is your current customer.
- [03:48]
A
Exactly.
- [03:49]
B
We have already sold them once. They believe in your vision, so, like, it's easy to give them another shot and like, see what else they would purchase. Anyway, so we're starting with this customer retention data, but the whole framework that we're going to use today, it's going to remain the same regardless of what dataset you're using. And for just a little bit more basics, I have created a customer retention folder under my Documents folder. So when we say work locally, we're working with basically files on our desktop. So what I did is in my Documents, I created a folder and I called it user retention codecs. And this file that I just showed you is actually saved in that folder. Now, what we're going to do is we're going to go to codecs. Codecs is basically think of like Claude cowork, but OpenAI's version of Claude Cowork and Claude Code all in once. I have both ChatGPT standalone apps as well as codecs. So in codecs what I'm going to do is I have already created a project. If you don't have a project created, you can just easily create project using this dropdown. So for example, right now you can say add new project and then basically it will tell you which folder and then you create the folder and then you will work within that folder. So in this case, I have already connected it to this user retention folder that I just showed you in my Documents folder. Now here, what I'm going to do is we're going to imagine a scenario. The scenario is your leadership is coming to you and it's basically 12pm on a Friday and they want to understand why user retention has dropped in the last week and what exactly is the root cause and they want an answer by 2pm so like you literally have two hours to like figure out which
- [05:39]
A
before would have been impossible, right?
- [05:41]
B
Exactly. And of course if you're going to present it to leadership, it has to be a deck. Like you can't just go with like some two word like summary. Like yes, that could work. But like if you really want to impress, like you want to put a deck together.
- [05:53]
A
I know we covered a lot today's show, but we've got some free resources that are going to help you take it to the next level. You can scan that QR code, you can click the link in the description below and get everything you need.
- [06:05]
B
So what we're going to do with Codex today, I'm going to give it a prompt and I'm going to say can you find what drove the retention drop basically last month? And I want you to build a cohort analysis and I want you to turn the top insights into a leadership tech. So it's a very simple prompt. Now we can obviously refine it once it starts developing, but this is the basic prompt I'm starting with. And the folder that I'm working with is User Attention, which has the CSV file that I showed you earlier is saved. Now in terms of the model, you have option to pick whatever model you want to pick from ChatGPT currently I have put it on Intelligence at medium and we're using GPT 5.5 and then all we're going to do is hit Enter and it's going to start working.
- [06:49]
A
What's interesting, it's like this seems very simple, but it's doing a lot of hard things. You know, you told it to make a cohort analysis, it needs to be a deck, which means it needs to graph that data in a way that's accurate and doesn't misrepresent that data. I think you're probably going to talk us too about how you make sure the data is correct. Right. Because I think one of the big challenges with any data analysis is like, am I sure my data is right?
- [07:11]
B
Exactly. So I think people have fear that the AI tools are not going to replace data analyst or data scientists. I think that's not true. As somebody who has worked professionally in the domain for the last 12 years, data analytics and data science is a lot more than just coding or doing analysis in Excel. It's a lot more about stakeholder management, applying the critical thinking and applying your analytical thinking. Because it's possible that AI is going to give you something wrong. And if you don't have the analytical thinking and the domain knowledge, you're going to give wrong answers and that's going to hurt your career more than is going to help.
- [07:45]
A
And as somebody who's like a real expert in data analysis and data science, like, what are the couple of tips, tricks, principles that if somebody doesn't have your level of experience, that they should just keep in their head when they're doing a project like this, first of
- [08:00]
B
all, you need to make sure, like the data that you are uploading and the tool that you're using, do you have the appropriate permissions from your company to be able to like upload that data into an AI tool? So the rule that I follow, yes, there are so many AI tools out there, but if you're using it for doing data analysis for your job, for example, at HubSpot, if you are doing like data analysis, like some marketing funnel analysis, what does HubSpot has subscription to at enterprise level? Is it Claude? Is it OpenAI? If that's correct, then use the internal tools that have like the enterprise security measures in there to use it for data analysis. So that's number one. That's bare basics. Don't take your company's data and upload it somewhere else. That's a big no. So that's definitely measurement number one. The second is when you're doing analysis with AI, have a clear problem that you're trying to basically solve with it. So unless you understand what you need to ask, what type of questions you need to ask, you're not going to be able to get a lot of value out of it. So like fully understand what you're trying to get from this data analysis and what the output you want to look like. For example, for my analysis, I want it to be a cohort level analysis and then I want it to be a leadership deck with explanation why exactly this retention basically dropped. Once you get the analysis, once you get the output, you need to spend a little bit of time looking at the numbers. Start with very simple. Does the math even make sense? Because sometimes AI tends to make mistakes.
- [09:32]
A
This is a great tip.
- [09:33]
B
Yeah, simple math. So if the revenue dropped by 12% but the customer retention dropped by 30%, that doesn't actually make sense. And you can only make that judgment if you have looked at the retention analysis in the past. So your experience, your background knowledge is going to help you quite a bit when you're working with AI to make judgment decisions and validate what AI is spitting it out for you. Right. So those are, I would say three fundamentals that you have to like keep in mind. Obviously like validate, validate, validate. Especially if you are just starting to use it. Eventually you'll build your trust with AI, but it will take time and if this is something that I'm going to put my name on it, I have to make sure this is actually the right correct.
- [10:21]
A
You just gave everybody 10 years worth of data analysis skills in like 70 seconds. It was, it was pretty impressive. Those are like everything you need to know to basically go from nothing to being competent at something like this, which is awesome.
- [10:35]
B
Okay, so we kicked it off and it's been working for the last five minutes. And what it's starting doing is first thing it did is looked at the data, it walks me step by step what exactly it's doing. And right now what it's doing is basically creating project files in the folder that we have given it and then it's looking for like appropriate libraries that it's going to use for analysis. It's looking for like permissions. So you may have to like babysit it a little bit to give it the right permissions. On Claude you will get like, do you give it access? And you have to like manually click allow access?
- [11:10]
A
Allow, allow, allow, allow, allow. Right. Like the, the, the two biggest challenges with whether it's Codex or Claude code or whatever, there's lots of them, but a few of them are allow, allow, allow and permissioning as well as just like managing everything locally and versus like the cloud and sometimes you need stuff in the cloud to share it. So you spend a lot of time manipulating files around different places, right?
- [11:30]
B
Yeah, exactly. So in the last six minutes it has primarily still looking at the right permissions and tools and libraries that it need. But we can jump to like a pre built analysis that I've done them analysis and I can walk you through in more detail.
- [11:45]
A
This is the magic of like the cooking shows on tv, right? It's like you're showing us what it's doing but now you're going to show us the kind of the end result that it was pre run. But that's an important thing for everybody to understand exactly is what you're showing is not like a 60 second project. This is I assume probably takes 30 minutes to like fully run.
- [12:05]
B
Yeah, so like we can see like the last one that I ran it took about nine minutes. So it went through everything, asked me a bunch of permissions and then it looked at the data in more detail. For example, it says I'm going to inspect the schema and data coverage next and why last month happened. And then it found that the data set basically it's still giving me description of the data. So it's saying Data runs from March 2 through May 3, 2026 with the latest week starting April 27 flagged as the dip period. So it's basically telling me more specifics in terms of the problem that I gave it. So and then it says that attention did fall by 66% in the prior week to 41% in the latest week. So basically it did identify that that attention did drop quite a bit. Now if you want to do like validation, this is where you would like quickly look at the data and like do some pivot tables and validate this number and see like if this seems correct. Okay, if this seems correct, let's keep moving forward. And then it said like it's looking at a couple of things and based on what it's seeing, it seems like mobile crash exposure jumps to 45% while email campaign exposure collapses to 1.5%. So it's kind of like hinting toward mobile but like it's going to keep going and then eventually after this finishes the analysis it actually does find that the pattern is now pretty clear. Customer weekly level retention drops from 72% to 46% and which is a drastic drop and it's related to new mobile version launch. So that's our root cause. So now the question that leadership was asking, why did last week drop? It's possible that the product and ENG team launched a new version of the mobile app and there was a bug and that led to a lot of crash outs and customers basically not retaining. Even though they were coming back on the site, they were not able to continue on the site because their session ended. Now it's not done yet. There is more. So because we asked it for a cohort analysis and we asked it for
- [14:06]
A
a PowerPoint, you have a robust data set here and that's really important. And it's possible if you're watching this, you might be doing something similar. And one of the things that it could come back with is like, I need more data. Right. Because you fortunately have a lot. So it can actually get a statistically significant conclusion from here. But it's very possible that if you're out there doing this, it's going to ask you for more data. So I think you want to start, I think with a pretty good size of data, more data than you think you need. Is that like a good thing that everybody should consider or how do you think about how much data to start a project like this with?
- [14:41]
B
Yeah, I would say this is again like where your judgment would come in because AI is going to just assume that it has limited space to work with. Like this is the only data set it needs to work with. So it's going to make conclusions based on whatever data set you give it. You, you give it like missing data is going to make conclusion based off that too. So that's where you need to figure out what exact dataset you need to be able to do the analysis. This is one of the mistakes that AI makes. It just assumes that it doesn't have the option to get additional data set from you. At least today. One of the things like data analysts, data scientists or anybody working with data struggles with is figuring out where the data lives. And a month ago when I was working at Google, I was actually able to use AI just to describe, this is the data set that I'm looking for. This is the SQL script that I want to write and pull the data. It would automatically figure out which tables the data lives in. It will automatically write the SQL. And finding the data set is one of the hardest challenge. When you work at a bigger company where there's so much data and it's like spread out everywhere, but even there's like an AI layer that is like all the rag, which is like your company's internal database information embedded in and you can quickly find where the data lives. So to your question, as an analyst you have to figure out what data is, the right data for this.
- [15:57]
A
You know that feeling when the strategy is done, the brief is written, everyone's aligned, and you realize someone still has to sit down and actually create all that content. That someone is you. And it's due tomorrow. Breeze assistant from HubSpot can help. It works right inside HubSpot. Drafting, campaign copy, blog post, emails, all, all in your voice, all grounded in your actual customer data. So you don't have to create content. You create content that converts. Check out HubSpot.com to learn more. Well, and I think what's fascinating about the show today is that it's like the ultimate example of like, hey, AI is really good and it's really helpful, but it is light years away from replacing humans. Everything you've talked about is you've talked about a step of, like, human intelligence, experience and discretion and judgment needed at, like, literally every step of the process. Right. And I do think there are some people who do kind of set it and forget it AI. And the AI is just going to figure it out and I'm not going to worry too much about it. And I think that's fundamentally a bad thing to do, but especially bad thing to do when it comes to data analysis, because there's so many gotchas. There can be hallucinations, there can be misinterpretations, and if it doesn't make sense to you as a human, it's probably wrong.
- [17:11]
B
Exactly. Yeah. I think Forbes wrote an article a couple of Years ago when ChatGPT operator came out and they basically made a very bold statement saying, like, data AI is going to replace data analysts with, like, ChatGPT operator. And I think that scared a lot of people. And I would say if you are a practitioner who has done data analysis, regardless of what your role is, you know that data analysis is way more than just crunching numbers. The way I treat it as like, let's say if I have an intern who is working for me and I know what problem I need to solve, I'll just give the crunching part and the coding part to my intern. And yes, intern will come back with something, but I need to, like, sit down with the intern and figure out what exactly makes sense, what doesn't make sense, what we need to devise and what we can just present as is.
- [17:56]
A
Yeah, I have this whole theory that no matter what your job is, there's like a magical creative component that the people who are great at that role have. Like, I think the Greek creatives, designers Editors, like, they know when to call something done versus, like to keep tweaking, right? They know this magic moment of when the art is done. I think for data analysis, for example, there's just like the best people at data analysis I've ever met. They're like the most creative question askers, right? They don't ask the obvious questions. They ask these counterintuitive, unique questions of the same information and get remarkably deep, different outcomes than anybody else who maybe just be a novice at that thing. Right. Do you see that?
- [18:39]
B
Oh, a lot. I don't know if it's like an internal inside joke or something like that. A lot of the times when like data scientists or data analysts have like stakeholders come to them, they will ask a question. But when you, like, dig in a little bit deeper, it turns out like the problem that they're trying to solve is completely different than the one that they came to us with. So you definitely have to figure out what the right problem, problem to solve is.
- [19:01]
A
Clarity of problem, and really creative questions around that problem.
- [19:04]
B
Exactly, exactly. You nailed it.
- [19:07]
A
Okay. That's awesome. I think what's fun is that we're building this analysis and kind of going through the skills and human aspects necessary to do that. And so you've obviously run this. You found, you know, in this sample data set kind of the core issues. But one of the things you talked about is like, if people remember a few minutes ago, it's like, oh, we need to present this to the executive team. We got to get a deck, we got to visualize that. So show me that component of all of this.
- [19:32]
B
Yeah. Okay. So the fun part, which I honestly, because personally I can write documents all day, tell me to write a six page document, I will write it for you. But as soon as you tell me to create a deck, like, I don't know, like, I am so lost. So one thing I personally love, that I can actually now create decks. Like take it from a CSV file to like actually create decks with AI. Now I will put a disclaimer. Codex doesn't create the prettiest decks. I would say, like, Claude and Gemini do better at Dex than Codex.
- [20:04]
A
I completely agree.
- [20:06]
B
But hey, it's a deck, so we can work with it. So after I did the analysis, it basically found the main drivers are the mobile app crashes. And then it basically gave me two deliverables. One is the cohort analysis that I asked it to give me. And then the second is the leadership deck. And you can actually see on the right side of the pane this is the cohort analysis that it did and we'll look at it in a little bit more detail. And then this is the deck that it created for me that I can present to the leadership. And then it also like gives you like coding files that it has created. But let's actually go back to the folder where it has saved all the files. So when we go to our documents folder now we can see like that it has created like bunch of folders and bunch of like. Basically here we'll see like it has created bunch of CSV files for the cohort analysis. And then for like deck it has also created like a bunch of photos and screenshots. And in the retention deliverables is the two files that I was showing you. One is the Excel file with the retention drop cohort analysis and the second one is the leadership deck. So let's actually look at the cohort analysis file and see what it did for us. Okay, so this is our source data that we gave it and it gave me a bunch of tabs that you can see here. So first tab is executive summary. So it gave me top leadership takeaways that you can see here. The number one takeaway is that the drop is considered on April 27th week, customer retention fell from 72 to 46. And then the second takeaway is like it's related to the mobile version launch and third is something related to email campaign. Now it has given me like the main metrics that the core metrics that it analyzed and what it found. But it also went deeper into like the analysis where it gave me like weekly retention trend. So here customer week retention is the primary KPI and this is the week that they started and how the active customer number has changed. So if you look at the weekly data, so like active customer in March, second week is 8:31. And then we keep going down. You can clearly see that the active customer number is 588. And then the retained customer drops significantly. And we also have our crash customer. So you can see like there's a 52.6% crash and so on. And then it also spit out a plot for us where it tells us like retention drops as we increase in exposure. So CDs 2 is our crash percentage and the Series 1, which is this blue line, it's our customer retention. So you can clearly see that they both are related event and one caused the other. Anyway, so here again like you will do the validation and you will go through the numbers, try to see like if it makes sense. Because otherwise, like, you can just present it as is. You have to like, do some math on your own to figure out like this actually makes sense or not. This is my favorite analysis that it created. So it basically shows this, the cohort analysis, where this column, column A is the signup month. It tells me what month customers signed up in and like how their retention has been. Yeah. Or weeks. And then you can see like April 27th is definitely. Retention is off for pretty much regardless of who signed up when.
- [23:29]
A
The interesting thing about this, as you're kind of walking through the actual Excel docs that's created. So one, Codex has built Excel docs for you and you can now go and view them, which is very valuable. And two, I think you're giving people real good insight into what data analysis is really about, which is like you're trying to find outliers and you know, in each of these different Excel tabs that you're looking at, it's like very clear. It's like, oh, something happened here. And you're trying to pinpoint when here is. And then what happened around that period of time to actually understand that you need to do something about it. And, and normally when you have a big performance change in a, whether it be a marketing team or your business, it's because you did something or because something happened to you that somebody else did. And you're trying to pinpoint what that is so that you can remedy that problem. And these are the types of Excel sheets that you need to look at. And if you're earlier in your career, you may not know what things like cohort analysis means or you're not super into data. One, ask the LLM and LLM will explain it to you. And if you want to work on your understanding, you can also just ask the LLM to create a dummy set of data and say, hey, I'm trying to learn how to better analyze these types of problems. Can you create dummy data and walk me through and help me figure that out? And you can learn that in real time and you can actually close a lot of the gap to expert knowledge like somebody like you has. But I think this is visualizing them in the spreadsheets a very helpful way versus just seeing like the couple line summary in Codex.
- [25:03]
B
Yeah, exactly. And then maybe you can like look at the spreadsheet and come up with like new hypothesis. For example, what exactly happened in April 20th week? Why is our retention so high? So like I would say like that's another analysis opportunity that you can like take back to your team and see like, can we investigate what we did in April 20th and can we repeat it more often? So it's possible that this may be related to another launch or something like that that drove customers back to the site. So the second deliverable that we had is our deck, which is basically what we want to present to leadership. So as you can see, like it put together how many seven slides and in slides basically is what we will take to the leadership and tell it like, yes, retention fell because mobile 4.3 version and these are the next steps we're taking to basically solve this problem. But for your 2pm, remember that earlier I said like we are working on a scenario where your VP or your director is like, can you explain why retention dropped? Now you're going to be able to like take this and not only share what you found, but also have a presentation if they want to like get on a call with you. And here you can explain that this is why customer retention dropped and it is related to the mobile launch that we did in April 27th week. And then these are like the three things that we have done like this to resolve things back to normal. Always go with what you're going to do about it. Don't just go with like the problem and the root cause. Always tell the leadership what exactly you're doing about it because that's going to be their next question and you want to be ahead of it. This is just my pro tip.
- [26:36]
A
That is a good pro tip.
- [26:38]
B
Anticipate the questions and then have the answers ready or prepare them in your deck. So if we look at this deck like as I said earlier, like Codex doesn't do a great job with decks. Like they don't look the prettiest. But for somebody like me who doesn't enjoy decks, I will take this and maybe I'll like modify some things like look through the data points that is spitting out, See if I want to add something or doesn't want to add something, then I will modify it. And here you can see there are some mistakes that it has made here. So I'll definitely go through it and fix it. But as like a baseline working deck that I can like take over and change. I think for that it already did a great job basically giving me something to start with.
- [27:22]
A
Yeah, it's a great starting point. It's really easy to make slides look better, especially with some of the other models. Claude and Gemini are great and I like your point of like getting ahead of the objections and you know, even that's like a prompt modification of like in the deck. Make sure you're including slides that would address any possible questions or objections. You know, a leader at the company would have would it would be a great way of doing this.
- [27:46]
B
Right.
- [27:47]
A
Okay. As we're about to close out, you have given us a masterclass on how to think about data, how to analyze data, how to solve problems with data. Is there anything else that people really need to know to make sure that they get this right? Because I want anybody watching this to say I can go do this and feel really confident about it.
- [28:05]
B
Yeah, I say like one thing that we briefly touched on but we didn't fully discuss is when you have your data, there is no such thing. Yes, we talked about some data is missing and we only have subset of data, but we didn't talk about the data that you have, how dirty or clean up it needs to be. So if you look at the codecs it started, it just assumed the data is already clean and there's nothing needs to be done here. If I were to redo it all over again, I will definitely ask it to look at what are missing values, what data needs to be cleaned. And obviously you're going to look at the data yourself too. But definitely follow that step because if your data has missing values or if it's incorrect, then the answers that you're going to get is incorrect. Anybody who's creating data analysis with AI just working under the assumption that the data is perfect, which in real world scenario it's never really perfect. So that's up to you to figure out how to get that clean data set from where. And if there is missing data, use AI to kind of like clean that up and normalize it before you do the analysis. Now I would say that what I showed today, it's not prominent today at all organizations. There are certain organization enterprise companies that have adopted AI. For example, when I was at Google, like we were heavily into using Gemini for data analysis. I know Meta is specifically pushing the PMs and basically everybody to become take over data analysis one of the skillset and use a Gentex analytics tool that they have adopted to like ask good questions and get an answer. I would say like five to ten years from now we're going to see more and more of this where this is going to become like a core skill set for anyone who wants to do data analysis or whoever touches data. So if you're learning this today and practicing this, you could not only be doing your work in a more productive way, but you could also teach other people in your team and in your organization how they can stay ahead of this and be the leader. So right now everybody's looking for, for people who like use AI to do your day to day job. So if you are that person, you could be that person at your company who is teaching other people how to do it.
- [30:14]
A
Yeah. Kind of your closing call to action is that this agentic data analysis is going to be a huge growth trajectory of career over the next few years. And if you're looking to kind of build your career or build the quality of company that you're trying to out there build, you need this as a skill set. And it's going to differentiate you out in the market because gone are the days of very traditional data analysis. You still need those fundamental skills, but the work needs to get done in this new and modern way.
- [30:44]
B
Yes. And for all my data analysts and data scientists and data engineers watching, your jobs are safe. They're not going away. Yes, they're going to change how we do things, but you're not getting replaced.
- [30:56]
A
You might be doing more and doing that in different ways and then there's an evolution of those jobs. But I think if one of my big takeaways is the human element of data analysis is more important than ever, not less important.
- [31:08]
B
Exactly.
- [31:09]
A
And that is a huge takeaway for everybody. You've given us a masterclass on how to actually get started and learn and basically get tough questions answered in a new and better way. Thank you so much, Sundas, for joining us. We really appreciate you. We'll see everybody next time on Marketing against the Green. I want to tell you about a podcast I love. It's called Nudge. It's hosted by Phil Agnew. It's brought to you by the HubSpot Podcast Network, the audio destination for business professionals and it's the UK's fastest growing business podcast. What I love about it is that the Nudge listeners love no fluff, no BS, evidence based marketing tactics they get in each episode. You're going to want to listen because this is like an MBA's worth of insight in every single podcast. And entrepreneurs, you're going to love the show because it's filled with repeatable, proven studies, not hearsay, not one off success stories. Marketers, you're going to love it because it discusses the psychology behind great marketing and what marketers are getting wrong. Listen to the Nudge wherever you get your podcasts.
- [32:12]
B
This data is wrong every freaking time. Have you heard of HubSpot? HubSpot is a CRM platform where everything is fully integrated. Whoa.
- [32:21]
A
I can see the client's home history, calls, support tickets, emails. And here's a task from three days ago. I totally missed HubSpot.
- [32:30]
B
Grow better.