Andres Glusman, CEO of DoWhatWorks, shares how his company's patented technology can help SaaS founders write more effective copy, identify website flaws, and learn from competitors.
What can you learn from your competitors' experiments?
Andres Glusman, CEO of DoWhatWorks, shares how his company's patented technology can help SaaS founders write more effective copy, identify website flaws, and learn from competitors.
Andres leverages his background in behavioral science to optimize growth for businesses. He agrees that a higher learning velocity leads businesses to find more efficient ways to allocate resources and achieve faster growth. Andres has a wealth of experience working with both startups and established companies, making him an ideal expert for marketing professionals aiming to augment their company's success.
What happens when you run A/B tests? Does it move the needle of tangible results or does it seem like you’re stuck in a race someone’s already finished?
For Andres Glusman, he saw this firsthand and thought of a way to do it better. It all began while running product and growth at Meetup. An astounding 80% of experiments failed to move the needle. This realization led Andres to co-found DoWhatWorks, a company that uses patented technology to help businesses improve their conversion rates by learning from the experiments of others.
In this episode of Markigy, Andres and Leanne dive into how to:
So what happens when a company's experiment goes horribly wrong? Or just doesn’t go wonderfully right? Find out in this episode of Markigy.
The actionable takeaways mentioned in this episode are:
To learn more subscribe to Markigy with Leanne Dow-Weimer.
This episode was produced and brought to you by Reignite Media.
[00:00:40] Leanne: Good morning. This is Leanne with Markigy, and I'm joined here today by Andres Glusman, co-founder and CEO of DoWhatWorks. Thank you so much for joining us. If you could just introduce us, tell us a little bit about
[00:00:55] Andres: yourself. My pleasure. Uh, so I'm Andres Glusman. I'm the, uh, co-founder, CEO of DoWhatWorks.
Boy, I can answer that in so many different ways. Um, you know, I'm a, I'm a behavioral scientist at heart and I'm really passionate about affecting human behavior and that manifests itself in my career by virtue of sort of figuring out, um, how to. Affect behaviors of people in business is specifically people by using experiments to help people, uh, help people do, do things that are in their best interests.
That's
[00:01:26] Leanne: great. And, and I love that you're a behavioral scientist, right? Because that's kind of like where our promise is, is that I think that marketing, marketing is a blend of science and creativity and thinking collaboratively. So I am having that science background probably helped you a lot, but could you.
Kind of help us understand, do what works and what it is and its story.
[00:01:52] Andres: Yeah. So do what works is a company that helps growth leaders do what works. We have an engine, that uses a patented technology to help, uh, that can detect the experiments that are being run by any company. And our clients use our products that is based on that data to be able to write more effective copy.
To be able to identify the parts of their user experience on their website that are maybe not working well. Uh, and to identify the things that they can try based on what wins and loses for other people that are in their space in order to be able to help them find wins faster and improve their conversion rates, which in this day and age is kind of one of the most important levers you can pull when budgets are tight and competitive.
Pressures are up, which we're definitely going through right now.
[00:02:39] Leanne: Yeah, absolutely. So there's so much there that is, is like, ah. Um, and, and before we started, I, I wanna warn people is that like I'm a kind of a fan girl about platforms like this, so I wanna be very clear, this is not like a sponsored episode.
I'm just a genuinely excited about this because as a marketer I hate wasting time. And, um, this is definitely a sore spot where, You know, I feel the pain point that this solves. Mm-hmm. Uh, so not trying to paint myself as the ideal client, but you know, um, I, I see the value in this, like, wholeheartedly. Um, now tell me a little bit about, like, you have this really cool example about QR codes and what you noticed.
Mm-hmm. Can you tell me about that again for everyone to hear?
[00:03:32] Andres: So one of the, um, and I don't want my, my part sound like an ad either, but, but I'll share with you an example here cuz um, we're seeing from our platform. So our platform has detected over 15,000 experiments and when we detect an experiment being run by a company, we classify them and, and organize them based on the kind of experiment we're seeing.
Is it a button? Are they changing button color? Are they changing copy? Are they changing imagery? And one of the trends that we saw in the last year, which is kind of fascinating, is we saw people start to experiment with using qr. On their website in order to try and help bridge the gap between your website and getting somebody to download the app on their mobile phone.
So companies, you know, have this age old problem, which is you're on a desktop computer or you're on your laptop and they want you to download your app on their app, on their, on your phone. So how do they get you to do that? You know, it's not the easiest thing. So what a lot of companies experimented with in the last year, which is really fascinating, wonderful big companies, uh, GoDaddy, Venmo, uh, the Robinhood, variety of different companies in their hero images, they started including or testing qr.
The idea was that you would aim your phone at the QR code on your screen. It'll give you a link that you can then tap to be able to download the app right there. And lo and behold, you're all set and done. And so, um, really interesting idea. And on the surface, that's like a really cool thing to try, uh, in this day and age, right, everyone now knows how to use a QR code.
Thanks to the pandemic. If you ever wanna order anything from a menu, you gotta do that. So we saw a lot of experiments where these companies were featuring QR codes on their log out user experience, um, for on, on key parts of their, of their real estate, on their homepage, you know, on their key sales. And what was fascinating about it and what we started doing is looking at, well, how often does that win?
And our engine can help us detect when an experiment wins and loses and it categorized it. So we looked at all of the QR code experiments and we just saw this fundamentally interesting pattern. I think you might already know the answer, so I won't you on it. Uh, but, but what we saw was that in fact it was almost universally the case that it.
Yeah. And yeah, go ahead. I,
[00:05:42] Leanne: I wonder if, you know, because I advise people and never put QR codes on their, um, digital. Uh, graphics because I'm like, they're already on their device. Like, you can't scan a QR code from your phone. Like, what are you gonna do? Like grab your, your iPad to do, like, that's crazy.
Like you don't, you're not gonna grab an additional device. Um, if you're on your desktop and you're trying to get a mobile app, that is the only place. Where I think it could even be relevant, but I wonder how much time they spent looking at the device usage and if they were able to only show it on desktop
[00:06:23] Andres: users.
Yeah, I mean, we're able to detect when the experiments are showing up with the platforms. And the results I'm sharing are from the desktop, uh, experiments. And I just think it's fascinating because it's one of those things where, um, two things are really fascinating. One, about a year ago, you know, one of the.
Like interesting Super Bowl ads was just a QR code floating around the screen. Yeah. And I think it created a lot of momentum in marketing communities to like, oh, we should do that too and we should create this thing. And what's fascinating a lot about in my own personal experience too, but is that things that are like popular and trending, Are popular and trending because they're popular and trending.
Not always because they work. Yeah. And this is one of those cases where you started seeing a lot of people sort of independently running these experiments and gloaming on to see if, uh, you know, to try it for themselves because they thought it was cool and they thought it was working for everyone else and they all sort of discovered the hard way themselves that, uh, nope, it didn't work.
Uh, pretty much for everyone who tried it in that instance. Now, I'm not gonna say the QR codes are never good. I. QR codes all the time. Now, at a, at a restaurant there's different instances where it's good, but the, like everything, there's nuance, there's a time and there's a place Yeah. For different components and, and different, um, levers and different approaches.
And what's really just like crazy interesting is that if you were to try everything, it would take you like a million years, you know? And it'd take you about a month to try anything for an average. And you know what I'm passionate about? Maybe, maybe this is what you're getting at with efficiency too, is how do you keep somebody from wasting a month of their time, you know, working with all the engineers to get that to happen and getting all the internal approvals and getting it through legal and then launching it only to see it not work.
That's totally fine. I mean, like, you know, there's a trust used for years that experiences what you get when you don't get what you want or learning as much you get when you don't get what you want. And, and, uh, that's certainly, uh, the case with a lot of experiments. Um, in fact, it's kind of shocking. Um, I was surprised to discover this, but it's consistent with my personal experience when I ran product and growth at Meetup.
Um, but is that 80% of experiments that people run do not positively move the needle?
[00:08:35] Leanne: I believe it. Um, I mean, that sounds more realistic than anything that I can imagine. QR codes I use when I don't have any better way to accomplish what I'm trying to accomplish. Hmm. Um, you know, I found myself in a role where I was ordering print publication advertising.
That is one of the few ways that you can try to create a guess as to how many people are engaging with your print ad. Um, or if you are doing like a bus. You know, um, those big, very top of funnel moves. Um, but, you know, there's, there's a couple things I wanna just kind of like pull out from everything you just said.
Cause there's so many golden things. Um, kind of thinking back to like, you know, the psychological premise of like mob theory and like, you know, how social proof works is that, Social proof doesn't always just mean I saw a testimonial on a website, it's, I saw other people doing this experiment and using QR codes on their website all over the place.
Yeah. And it's amazing how powerful it is. Um, even when we
[00:09:51] Andres: don't intend it to be. Yeah. One of the, uh, funny things I witnessed is, um, there's a, a company that is very, um, I, I saw one company running a test. I, I don't even remember what they specifically, they were testing, but it was an approach to, uh, to getting people to sign up.
And then I saw their competitor copy the variant of the test that they were just running. So the competitor was like, oh, they're doing this thing. I need to do exactly what they just did. They saw that, they were like, they actually happened to witness the losing side of that variant in action when it was live.
And so then they went ahead and copied them and they reran that experiment themselves, uh, just because they were just following what other people did. And so, look, I, I think, you know, we all stand on the shoulders of giants, like the, this industry is what it is because we've been able to learn from each other.
Get inspired by other people. Um, the art and the science is really figuring out what to copy and what not to copy or what to use, the inspiration and what to leave behind. And, and to your point, you're a hundred percent right. I think people get a little too carried away that, oh, this big brand did, apple did this, therefore we should do it.
Yeah. You're not Apple. And there's a lot of reasons why Apple's very successful and it's got nothing to do with, you know, wearing a black turtleneck. It's got very little to do with wearing black turtleneck, but people sort of copy the wrong parts of it. I, I didn't mean to offend you
[00:11:12] Leanne: if you're only listening.
I'm wearing a black turtleneck. But, but simply because I like to wear black shirts on this show so that when I'm recording video, um, my pink cheeks, I can color balance correctly. Um,
that's
[00:11:26] Andres: fantastic. Uh, that's really. My mouth there.
[00:11:31] Leanne: No, it's hilarious. Um, but, but it's true. You know, I'm not wearing a black turtleneck cuz I'm trying to emulate the founders that wore black turtleneck.
I just so happen to be, you know, early on West Coast time and I'm like this, like mine is more functionally based. Um, But kind of coming back to the, our topic at hand is I have this concept I call the learning velocity. And what it means is basically that like pipeline velocity an organization, the faster, the higher their velocity of learning is, the more efficient they are with money, the more they're gonna grow and they're really gonna see those business results.
And where, what I tell people to look for is to look for those places where you. can Um, have your starting point in that velocity curve higher. And this is one of those places where instead of starting in, you know, the bottom left, like very zero corner, um, like a serial entrepreneurial, that's like a person who's failed a few times.
You don't start in that bottom corner. You have a higher starting point to your curve because there's already the learning that other people.
did
[00:12:44] Andres: Yeah, I love that. I love that. It almost reminds me of, um, you know, um, the rocket ship or the, the, the, the people call it hockey ship. Hockey stick. Yeah. The hockey stick.
The hockey stick. And what's, what's actually happening? So, right, you're, you're working, you're working really, really hard, and you're trying, uh, for as long as you possibly can, and you're doing everything you can to try and get something to stick. And no matter how much effort you're putting in, you're not necessarily seeing the.
Because every, every one of those things is getting you closer and closer to, to, to what actually works. Yeah. And it's upon discovering what actually works, that you start to bend upwards and you start doubling down and doing more of the things that are like that. And that's why you start seeing people hockey stick upwards.
And so to your point, the learning velocity, like you're talking about, It determines, well, how long are you on the flat part of
[00:13:28] Leanne: that curve? Are you gonna make it, are you gonna run outta your cash runway?
[00:13:32] Andres: Big time. Big time. Right. And how quickly can you get to the bend in the elbow? And, uh, and that's perfect.
You know, one of our, um, don't steal it. Yeah. Well, one of our, one of our clients, which, uh, which I did steal, uh, says, you know, sort of like, uh, well, your goal is to sort of skip the regular season Yeah. And go straight to the playoffs. Yeah. You know, so you can use data. You know, he says he, he worked with us because.
You use our data to skip the regular season and just go straight to the playoffs where things get interesting and, and to your point, that's like how, how do you start as close to the bend in the elbow as possible so that you're still learning, but you're just not turning over a lot of stones that you don't need to be turning over.
Yeah. You're
[00:14:08] Leanne: learning in a more relevant way that is very on target with your goals and your audience. That's right. Um, because, you know, From what I understand is that you're able to, you know, you're comparing against like businesses on your platform. Mm-hmm. And, and there's some benefit, there's a lot of benefit to that, but you wanna make sure that your brand in your specific setup is formulated in a way that is unique.
And so you can't always just take what works from someone else. That's right. Otherwise, you're, you're
[00:14:44] Andres: just everybody. That's exactly right. And you wanna be very mindful of, and this, this, again, there's nuance here, but you wanna be very mindful, like you're saying, are you, there are certain behaviors that make a lot of sense in a business context that don't make sense in a consumer context.
Yeah. Or vice versa. Um, and so the things that work in a, in a B2B setting maybe are not the things that, that streaming companies should be using. Uh, that being said, the coolest thing is like the places for innovation. Is when you jump. Mm-hmm. And you look at areas that have nothing to do with your, that in theory, they're not competitors, but you can learn a lot from.
For example, Southwest Airlines is super famous for being very efficient at getting their planes off the ground. They make all their money. When your plane's on the ground, you're an airline, doesn't make money and they want to get the plane back up and running as fast in the air as fast as possible. And what they then do and need to do is have their, their ground crews be really efficient.
They study the, they don't study United Airlines. They don't study Delta to figure out how to do that faster. They study Formula One pick cruise. Yeah. Right. And they use the insights from Formula One pick cruise. To figure out, okay, what can we do with our setup to get our planes back up, you know, back up in the air as fast as possible.
And so in a similar way, there's an opportunity to learn from the people who are doing, uh, nothing like what you're doing. You know, if you're a B2B company, maybe you could learn from direct to consumer companies, uh, and steal what's working for them as opposed to just learning from evaluating the experiments and the things that your competitors.
Yeah, absolutely.
[00:16:19] Leanne: You mentioned that, um, B2B customers also now have B2C level expectations, and that's a, that's a tenant that like I bring to everything I do is that, you know, you do need to blur the lines instead of just staying in your vertical because humans are humans. Mm-hmm. I don't show up with a com.
Uh, I show up with a tempered personality at work. Um, I act a little bit more professionally, a little bit less casually, but I'm still me, like, I'm still gonna be like, wow, this customer service sucks. I'm not putting up with this. I've got other things to do. I'm moving on and we are picking a different vendor,
[00:17:02] Andres: you know?
Yeah. It's, it's one of those things where, um, it's probably one of the best things that's happened to software, uh, in, in the last, uh, decade or so, is that as you know, SASS has really taken off the end user experience has become that much more important and getting people the, to adopt things on their. As opposed to just having it be sold to a, to a CIO somewhere and kind of forced into the organization having it b bubble up.
And that's really raised the bar on the customer experience. On the consumer experience. On the user experience. And rightly so, the, the expectations have gone up and, and, and so if you're creating a brand new software in this day and age for B, the B companies, You have to approach it in a way that, um, and be of the standard that a Netflix is gonna be at that uh, you know, that a Spotify, that a Hulu, that a Disney, whatever the case may be, you need to sort of deliver on a similar like experience, um, as best as you can because that's what the expectations are gonna be.
[00:17:56] Leanne: Yeah. It comes back to your standing on the shoulders of giants, you know? Um, it, these things that were a big deal 10 years ago have just become pillars of our. It's true. And that's so cool because, you know, it, it allows people right now that are developing something innovative that are develop, that are, you know, those next garage startup people in 30 years, what could they be?
You know, it's only gonna grow. Um, and tools like this will help it there. Um, you know, that being said, if you totally. Are inefficient and bad at business, like it's probably not gonna grow. Um, but it's true. Yeah. There, that's the thing about entrepreneurialism that, that really gets me excited is just the possibilities.
The, the what ifs and the dream and the inspiration that, um, the visionary leadership aspect of it, that, of the possibilities. Um, So kind of thinking about when this is a good fit, we talked about how you, you do have to be nuanced and you do have to kind of be mindful with how you use it. When is it not appropriate for someone to kind of try your platform or, you know, who's this not for?
[00:19:16] Andres: Yeah, so, um, there's kind of two main parts of the platform. There's the, there's a part that helps you write great ad copy by using AI and the winning variance from other. And then there's the part that is, uh, helps you get people through the front door. So get people through the front door and through the front door.
Um, when it comes to getting people through the front door, um, the reality is that, um, if you don't have a ton of traffic coming to your website, so if you're early stages, if you're an early startup, you know you wanna do the best you can to get the best version of a website up and running. Yeah. A 10%, 15%, 20%, even 30% improvement in your conversion rates on a very, very, very small number may not be worth all the effort that it's gonna take in order to optimize your website to, um, find and squeak out and collect all the wins to get that improvement.
So the number one thing that you need, uh, in order to take advantage of a platform that helps. Improve the conversion rates on your side is, is traffic. If you don't have traffic, if you're not exposed to people, then uh, then the impact will be relatively small and there's maybe bigger wins you can get by focusing that same resource, the dollars, the energy, uh, on other, uh, like creating a good solid baseline experience, but then focusing those resources maybe on getting better distribution or getting other things.
Uh, typically at this juncture, we work more with mid-market and enterprise size companies. We work with six of the top streaming brands like b2b, sas, unicorns. Those are the folks who like, love, love, love us because when they get a win, it's multiplied times, you know, billions of dollars. That's kind of the big deal.
Uh, when you have a win that's multiplied times, tens of dollars or hundreds of dollars in a day, uh, maybe that's not as important, right? So number one is just that conversion has to be something that's meaningful for. On that side. Uh, that being said, we are launching a brand new ad product, uh, that is um, allows you to make better optimizations on the copy you're running, on your ads and search engine marketing and then that is good for anybody who's running ads cuz basically every single percent improvement you get on your clickthrough rates.
And your engagement basically gets you more volume for your money and like a better quality score and therefore more love, right? Yeah. And so, uh, we did actually, part of my motivation is cuz I'm an entrepreneur and I love working with the entrepreneur community that the, that new product we're really excited about because we can now actually start working with, uh, you know, our brothers and sisters in the startup community and, and start helping, uh, helping them, uh, you know, get more customers and get more efficient, as you say, with the money they're.
[00:21:57] Leanne: there's just so many, so many ways that like, you can approach this, you know, and I think that you called out something and that I, I wanna reiterate this to the very early stage people, is that sometimes mediocre is enough. Hmm. You know, your website's number one job at the very beginning is just to prove that you.
It is to answer questions. You know, your, your basic FAQs, it doesn't have to be glamorous. It has to exist. Mm-hmm. Um, and you can, you can, you know it, you wanna do it as best you can, but mediocre, not, not bad, but mediocre can be enough. And, and when you really have very few, you know, if you have a marketing department that's like one person, then you have bigger fish to.
[00:22:49] Andres: Um, as long, as, long as it meets like a professional standard, it can't look, it can't look like, uh, like a middle schooler put it together on, uh, on geo cities, but, which is not what we're saying. I know. But as long as it looks like, uh, like it, like it's professional and it conveys enough about what it is that you need conveyed, um, obsessing about squeaking out every single pixel and making, uh, small improvements are not necessarily gonna move the needle in ways that bigger Boulder changes would, would do at that ju.
And that's what you need to be testing because you have so little traffic that that's where you need to be focusing your resources.
[00:23:22] Leanne: Yeah, exactly. It's about moving the needle. Um, and I could definitely see how this could some help a company that is in like mid-market, move up market. Mm-hmm. Um, and and things like that.
So, you know, because they have the, the investment and the resources to make it actionable. That's right. Cause.
[00:23:45] Andres: It's actually really, and they're interesting because they don't yet have the traffic that their bigger competitors do. Yeah. And so they have a little bit of a resource. They, they, they, they, they have the footing that the beachhead, uh, but they really need to figure out how to increase the velocity and their traffic limits, their ability mm-hmm.
To get experiments and get learning. And so if you can 10 x the. Experiments they're learning from on a given month, well then they've been 10 X their velocity and they're gonna find the good stuff faster. They're gonna find the stuff that works sooner.
[00:24:17] Leanne: Yeah, there's, there's a lot of cool stuff, um, with AI that, um, and things like this.
Let's talk about the future of marketing. Hmm. Where do you want it to go? Verse where do you think it should, or where do you think you want it to go? Verse where do you think it's going with or without our opinion?
[00:24:36] Andres: Yeah. What a great question. It's such, um, you, you just referenced AI too, so obviously that's gonna be a big, big part of it. But where's the future of marketing going? We are in a world right now. Um, that's getting noisier and noisier. And what's funny is, is if I had had this conversation with you 20 years ago, we probably would've been talking about how noisy it is with the internet.
Right? And now it's even like a thousand times noisier. And we probably would've been having the exact same conversation 20 years before that with cable television and how it's getting noisy now. But it's really truly a, a noisier and noisier world. It's harder and harder to break through, and that's only gonna get worse.
The easier it gets to produce content and the easier it gets to produce words and put words out there. And so, um, the value or the impact that you need to have from, um, the attention that you're able to garner when you're able to get a little bit of attention, like harnessing that and capitalizing it, is gonna become so much more important because it's gonna become so much scarcer.
Mm-hmm. And, um, it's really interesting in that it's a good thing. That AI is gonna make it possible for us to like 10 x our productivity or a thousand x our productivity. With regard to that, we're starting to use the AI on this new copywriting tool, uh, based on the winning variance that we see people run.
It is so cool. I mean, it, it kind of is like, it's magical. We didn't expect AI to be this good this quick, and it's, it's gotten there and so you can only imagine how much better it's gonna. Five, 10 years from now, it's gonna be a little scary. Um, but all that being said, what it means is that like things are gonna become very efficient.
And what I think the challenge is gonna be is how do you bring the humanity through? How do you, like, convey the soul of what you're working on and make an emotional connection. And the do that in a way that breaks through the, the, the, the things that are written by robots essentially. So I, I do hope that humans will be able to suss out and sort of get a different feel for what's written by robots.
But the reality is, um, that the parts of humans are going to become more of an editor or more curator than creator. And it's going to be the role of the marketer to be like really good at figuring out how to like generate the right. stuff And then curate it, and it's going to be a curation job more than it's going to be a production job.
[00:26:59] Leanne: Absolutely. I think one of the super skills that marketers should develop right now, yesterday, 10 years ago, is learning how to intimately understand the nuances of their customers by directly engaging with them, by watching social platforms, by understanding sentiment analysis, um, by really looking at everything from that macro level systems approach.
Um, and looking at what the feedback loops are interacting to either like, make something bigger or smaller or amplify things. Um,
[00:27:37] Andres: you're totally, I, yeah, go ahead.
[00:27:39] Leanne: W when I say this, this is a, a tangent, so apologies, but I'm gonna do it anyways. Um, if you look at what I call the panic feedback loop, and what this is, is, We as humans, when we see something alarming, it gets and keeps our attention.
Mm-hmm. Because that's a survival instinct. That's how we're hardwired. That is necessary, right? Mm-hmm. Um, and how this has been exploited is the, if it bleeds, it leads media cycles. And so it creates this loop of, you know, first, you know, something alarming will happen and someone will be alarmed, and then they'll like shout it to everyone they know.
And then it gets, it starts to go viral on the socials, and then it becomes a news story, and then the feedback loop amplifies and amplifies and amplifies, and all of a sudden everybody's panicked about whatever that may be. And so when you look at it, when you're able to take like five steps backwards and intersect where people, how people operate, how the platforms operate, and how, you know, the big corporations for media operate and understand what's in each of those players' best interests that they're going to do and how they intersect.
Mm-hmm. You're better able to humanize and understand and amplify. Your stuff. Now, you should use it ethically. Please don't create panic feedback loops unless there's like a real reason to panic.So, um, humanizing it and, and getting to know how people operate and how, how your individual people operate and all of.
You know, 360 holistic sense of them I think is, is gonna be really, really impactful for, you know, like if you mentioned like the curation of, of how your, your entire marketing
[00:29:32] Andres: flows. You're, you're totally right. Right? Because what, what is the, the AI is gonna be great at generating stuff. Mm-hmm. But it's only gonna be as good as the insight that it uses to generate the.
So AI is, if you have a great training set of data, if you tell it what to look at in the right ways, it'll generate really good stuff. But if you tell it the wrong things to look at, it's gonna generate things that are just not right. Right. Yeah. Yeah. To the point, the insights, the, the skill of knowing what your customers want and knowing the components and knowing what they're gonna respond to and the, the, the what works and doesn't work and their, and their specific nuance is, so I.
To the training data because it's garbage and garbage out. Yes, exactly. You can use your smart insight there. That will be a fundamental differentiator in a world. We're making things as easy. Yeah.
[00:30:30] Leanne: Right now, the way it stands is that we tell the ai, we train it, we teach it, we request it. It is just a code of logic.
It is not synthesizing stuff. It is. just Following instructions and if you write bad instructions, you're going to get bad results.
[00:30:48] Andres: That's right. That's right. It's only as good as the training set. And that's what's really neat. I mean, that's where I think that we're excited about that because, um, we can't wait to see what our training, what our data helps to train it to do.
Yeah. Uh, but, but it's true across the board for any kind of problem that you want AI to solve, which is fundamentally knowing what data to. Is what will ultimately determine the outcome of whether the thing it gives you back is any good or not?
[00:31:14] Leanne: Yeah, I can, I can already think of some misinformation that people have created with chat G P T because they fed it like
[00:31:22] Andres: it was easy.
It's so convincing. It's really, really convincing. And it's wrong. I mean, it will get better, but again, it will get better because the training will get better and the, the rules around sort of like refining what it comes back with will get. But only as a result of the cycles and the iteration loops, um, based on the data that's put in.
[00:31:39] Leanne: I, I just wanna keep talking to you. You have so many, like, great insights and knowledge. Um, I'm gonna ask a bonus question, uh, that I didn't warn you about. So if you think of one of the coolest things besides this that you've done, uh, professionally, Hmm. What, what would it.
[00:31:56] Andres: what am I really proud? Like when, when I think about things that I look back on my career and, um, what I'm really proud of are. And what I really love and what motivates me is like innovation and breakthroughs. And so anytime where I'm able to feel like I created a breakthrough on something, I rewrote the rules on something pretty fundamentally important, um, to find a like saying, yeah, that's awesome.
Conventional wisdom is cool, but what if we did this? Mm-hmm. I've done that a lot and it doesn't always work, but when that works, it's so cool. And, and so the times where I've done that, there's just like many, many times where it worked out, which were, were great and around, you know, getting involved and running some of the first experiments online or building a usability lab that sort of threw out all of the rules of usability testing and for the same amount of money for like under $32,000 a year.
Instead of running one day of usability sessions, be able to run, you know, 400 sessions in a year. Uh, you know, or completely just rethinking the end-to-end experience to generate 10 x the results in, in kind of same amount of costs or less. So that's the kind of stuff that I'm really motivated by. I can share specific stories, um, but, but, but what a, what a cool question.
So, to me it's, it's anything I can do that's just like a breakthrough, um, by just. Going to the first principles and rewriting the rules. Uh, I'm, I'm an immigrant, I'm a youngest child. So these are all like factors for me that like,
[00:33:27] Leanne: yeah, rules, rules need not apply.
[00:33:29] Andres: If, if you don't, uh, if, if you're the youngest and you follow all the rules, uh, then you're gonna, you're, you're, you're eating last.
But, uh, but if you can rewrite the rules, um, then uh, Then, you know, uh, you have a fighting chance. And my siblings were wonderful, so, so don't, don't get me wrong, they're, they're wonderful, generous humans. Uh, but I think as a youngest child, it's pretty consistent that that's the kind of behavior you'll.
[00:33:49] Leanne: Yeah, I, I'm the youngest too. And, and I like to think that youngest children we're, we're scrappy. Right? Okay. Yeah. Yeah. You know, um, we see maybe there's an older sibling that, uh, you know, is very, that, you know, the first child that like, you know, follows the roles and parents feel like, you know, they, they, they know what they're doing and then all of a sudden we come along or we're like, Nope.
We're gonna do this.
[00:34:10] Andres: Very much so. So I love that. I love seeing stories about that. I like doing that. I love, I feel pride when I'm like successful at taking something and turning it on its head. Um, because you can, yeah. That's, that's really fun for me. And it helps because I'm a youngest and because I'm a foreigner, uh, you just sort of see things as more arbitrary, um, which, which I think is, is, is a nice thing.
Awesome.
[00:34:35] Leanne: No, I, I definitely appreciate that. And then I, I don't wanna, you're such an amazing resource that I don't wanna like underutilize my time asking you questions, so I'm gonna throw you another one. Uh, it'll be the last one, I swear. But what is a question that you wish people asked you but they, like, they miss out on asking you?
What is an unasked question that we should be asking you to
[00:34:59] Andres: learn? Um, really, really that first of all, that's a great question. That's the question. Um, I stole it from someone. I'm sorry. It's a good one. Um,
what is the unasked question or what, what, what is the thing that I'm passionate about that I could share that I, that or learn? Um, Or if you're solving a
[00:35:17] Leanne: problem for somebody and they're asking, you know, like 10 questions and you're like, no, what you really should be asking is this. Like you're missing the point.
[00:35:24] Andres: All right. I will give you my favorite question in that regard. So it was almost, um, a caricature like I was made fun of at Meetup because, um, I would ask this every single time we were embarking on anything we were about to launch anything. The thing I would ask of my team every single time was, what has to break in your favor in order for this to work out?
Yeah. What's the number one thing that has to break in your favor in order for this to work out, and that changes over the course of any given project. The, the thing that initially has to break in your favor is that maybe you have to know, like, like we're launching a new company or working on a project that you know your hunch as to what the problem.
It needs to break in your favor. People actually have to have this problem. They have to actually wanna improve their conversion rates, or people have this problem, they actually wanna lose weight. That's not those, those aren't hard things necessarily to, to, to do the question. Then the next question was, oh, well then the thing that has to break in my favor is that the way that I'm thinking about it has to align with how they think about it.
And then my solution actually has to work, and then I have to be able to get in front of people or I have to do this. That there's whatever project you're working on, if you ask the question up front, what has to break in your favor in order for this, That reveals all of the stuff that is not important and lets you hone in on the one thing that is like the number one thing.
Yeah. And what you focus on right away and nothing matters until you solve that. Yeah. And then once you saw that, it's like going to an amusement park where you only get a ticket to the next ride once you've completed the. So if you've solved that one thing that has to break in your favor, you get rewarded with getting to answer that question again.
What's the next thing that has to break in your favor? What's the next thing? What's the next thing? What's the next thing? Over and over and over again? And, uh, it that's a methodology that is extremely, as a framework goes, is extremely flexible. And I will be hard pressed to find a situation where that's not gonna work.
Maybe the only place where that wouldn't work is if, is it maybe in a two-sided marketplace where two things have break in your favor at the same time, but that's the thing that has to break in your favor is the two things after break in your favor. So it's a very flexible framework, um, and it's the most useful thing that I've used in the times where I've been most successful.
Awesome.
[00:37:48] Leanne: I love that question. And you know, the project management like Brain In Me comes out with, uh, in order to mitigate risk, you just flip it on its side. You say what can't break in order for
[00:38:00] Andres: this to work. That's exactly right. That's exactly
[00:38:03] Leanne: right. Uh, awesome. That's gonna make my whole day. I love
[00:38:07] Andres: that question.
I love it. And it's, it's bright and early for you too. So you,
[00:38:15] Leanne: it's, um, what, so. Thank you so much for sharing all of this, and like I said, you're an incredible resource and you're just, you're brilliant. And so I'm really thankful to be able to have this conversation with you. Um, if someone else wanted to reach out, AI and, you know, connect with you, what are some ways that they can do that?
[00:38:34] Andres: Yeah, so the best place is through LinkedIn. Uh, that's where I like to share, uh, whatever's on my mind. Um, but they can also, LinkedIn is just, you know, Glusman, uh, Andres Glusman linkedin. The, um, you know, I'll put my email out there too if you wanna to get ahold of me through email. It's just andres@dowhatworks.io a n d r e s do what works.io. Awesome.
[00:38:53] Leanne: Thank you for spelling it out too. Cause that was gonna be my next question. Uh, well this is Leanne from Markigy and if you guys have any questions that you wanna ask me, uh, just reach out to info@markigy m a r k i g y.com and thank you so much for listening and thank you so much for being here.
[00:39:14] Andres: It was my pleasure. Such a fun conversation. Thank you.