The Business of Tech: AI is eating market research

WEBVTT - The Business of Tech: AI is eating market research Welcome to the business of tech. I'm your host, Peter Griffin Market Research. It's traditionally been a slow, expensive game, the exclusive domain of big agencies with big budgets. But what if you could get the same quality of insight that used to take six weeks and tens of thousands of dollars overnight for a fraction of the cost.

That's exactly what today's guest is building. I'm talking with James Donald, CEO and co founder of Ideally, an Auckland based startup that's using AI to fundamentally reimagine how brands listen to their customers, and James has a lot to be smiling about right now. Ideally just closed a sixty million dollar Series A rays, valuing the company at one hundred million New Zealand dollars. That's a significant psychological threshold for any QWI startup, and Ideally is in very good company.

In just the last month, two other AI centric New Zealand startups have now hit that one hundred million dollar valuation mark. Physical AI company Antioch and Caruso, and Auckland based platform using AI to help fund management in private markets. It's a genuine moment of momentum for New Zealand's fledgling AI ecosystem. Now ideally operates in what James calls a forty billion dollar slice of the global market research industry, and right now ninety percent of that span still goes to people heavy agencies like Cantar and Nielsen, the big players, the names we know well.

The AI disruption of that world is well and truly underway, and Ideally is at the sharp end of it. In the conversation, we dig into some of the big questions reshaping the research industry. We'll talk about the rise of synthetic data and why James believes that for the decisions that really matter, real human data is still hard to beat. We'll explore what Ideally calls living data, the idea that market intelligence shouldn't sit in a forgotten PDF, but should be a constantly evolving, always accessible understanding of your customer.

And we'll get into the SaaS apocalypse, that moment earlier this year when hundreds of billions were wiped off the value of software companies in a single week, because Ideally is very much part of that story. A new wave of AI native platforms looking to steal market share from the expense of incumbents and put powerful research tools directly into the hands of the people who know their business best. This is a story about what happens when you combine real human data with AI, and why that combination might just be the future of how companies make their most important decisions.

So let's get into it. Here's James Donald from Ideally. James, welcome to the business of tech. How are you doing.

Yeah, I'm doing very well. Thank you. I've been on a bit of a trip. Well, you must be on top of the world.

You've just raised what sixteen million dollars, valuing the company Ideally at one hundred million New Zealand, which is a bit of a psychological threshold for I think for Kiwi companies, and you're in a very good company. In the last few weeks to other sort of a centric New Zealand startups have also raised a lot of money, valuing themselves at one hundred million dollars. Antiocha company I featured recently, a sort of robotics and physical AI company, raised a fair chunk of money.

Caruso another Auckland based company that's I think they make software for managing private markets. That's a massive market. You know, you've got the public listed companies and then you've got all of these companies where there's private assets, so AI for that so well, it really seems like there's a lot of momentum building around New Zealand companies that are using AI to really good effect. Yeah, isn't that awesome.

There's some highly capable people here in New Zealand with a global mindset. And what was it like for you? Do you enjoy the fundraising side of things? Enjoys a loaded word.

I got a lot of value out of it. Yeah, I did enjoy it. There's some pretty clever people you get to interact with as you're sort of sharing your business and sort of uncovering the mechanics for artworks and all sort of ticklating the vision. And they're not afraid to give feedback.

And so in some senses, it's quite fortunate to be able to put that out to lots of different people. Now, you don't listen to necessarily have to listen to all those different perspectives, but certainly when you start to see some themes come through, then you actually, there probably is something here that I need to be thinking out for the business and so we were lucky to sort of have a range of different investors who are interested in the business. But one of the big take ways was actually helping to sharpen actually what the next phase of the business needs to look like to help keep positioning it for that sort of ultimate goal of becoming more and more valuable business with the eventual exit, which is sort of all investments BAT companies need to be thinking about.

Yeah, so this obviously was a Series A investment, so you're definitely on that pathway. Some great companies involved, Shearwater Capital, who were one of the founding investors in Wise Tech Global, a great company. You've got Altered Capital, Ice How's ventures of course have been behind so many great New Zealand startups. So on that next phase to really scaling up this business, you've got a lot of great advisors on board.

Yeah, that was sort of part of our criteria for sort of how we want to think about choosing those partners, because it's not just who's. Going to give you money. What I'm really looking for when we're looking for a new partner, what more can they bring beyond just money, experience contacts both for talent, the people that we might want to hire in the business or some pathways into the US been one of our growth markets, and so that's sort of how we ended up with a mixture because all of those different parties have different benefits and things that they can bring to the table just to help us be the best we can be to deliver on this opportunity.

I think what we're saying really justifies what I'm hearing from people in the market, both in the startup community in tech but also in the investment community that while we've had constrained capital in New Zealand, traditionally, there's actually enough money there if the opportunity and if the company and its founders are really good, and we do have a lot of really good companies and founders at the moment, so they're attracting that capital. So I don't think the constraints so there to the extent that they have been.

It's obviously cyclic, and we've seen a couple of tough years in VC there's been less money, but it seems to be coming back. Yeah. So I mean one of my perspectives of sort of raising money in this round is there is lots of money available, it's just they're pecky and how they want to deploy it and choose it. It's funny talking to some of the US based investors.

They were like, normally, when you're growing a double double every year, very investable opportunity. But AI has actually been inflating almost expectations that there's sort of a cohort of them looking for three, five, ten x growth even at this stage. I mean, look at Lovable and so that performs some set of expectations. But here on this side of the world, I mean it's a big raise.

And that's actually where we've got multiple investment partners. I was chatting to someone recently because they just sort of had the expectation that a lead investor might be filling eighty percent of the round, but in this case to be able to raise the size around out of. The side of the world. That's one of the reasons why you're sort of adding a few different partners because they're each and in a way there allocation for this type of stage in phase of business, and so you need to add up a couple to get sort of essencized war chest to go and go and sort of compete with the best in the US.

I think it's great, you know, because you get as you say, they've all got different approaches to VC and investing in different skill sets. And what we've seen was, you know, the best out of Silicon Valley often is they have two or three companies that have a significant stake in the company, invests significant dollars and you're getting the krem Della cram across you know, the industry, So you're not just getting one perspective. So I guess it's a risk mitigation strategy for them, but also for you to some respect as well.

Yeah, very much. So it's not your first rodeo James on startups, Well, well we'll go through that. You've you've been with leapbooking and travel software, then Yonder, a great company that that exited from New Zealand to us. A company bought that company.

But let's just go back a little bit further to your origins as an engineer and as an oil and gas guy, which I was really intrigued to find out that over a decade at Shell a little bit with Todd back here looking in the Taranaki region at oil and gas assets. So maybe chart your journey through university and into oil and gas. Oil and gas was serendipitous. As a good student, you're going to the scholarship's office and looking for ways to find some money to help fun for your studies, and its oil and gas.

I got a scholarship, I had to go work for Shell during the summer, and that's sort of where I got exposed to an industry I didn't even know existed in New Zealand at that time. I thought it was just a bigger lake under the ground. And when I finished the university, because I also did a commerce or economics degree on top of the engineering, and I knew I wanted to go to the business direction I management, consulting or while on gas or an engineering route, sort of knowing that eventually the two will converge.

Moving to Norway straight out of university was a pretty compelling option for a twenty one year old, so I took that route. But then, after ten years and sort of a declining industry, I just I could just see the writing on the wall that I needed to be controlling my own destiny, and technology just looked like such a great environment. It was ten years ago now when Airbnb. Was kind of in their scale up stage, and why Combinator was really in its heyday.

Were Sam Ultman still there, Paul Graham, a bunch of the legends in Silicon Valley, And I just got quite inspired at that point in time. Had another friend, actually, Jonathan Good, to go through Why Combinator with his own business, and so I just got inspired by that San Francisco mission focused technology solving problems. So I'm like, I can do that. But that's when you mentioned one of my first businesses that failed very fast in like three to six months.

It was a two side marketplace. One side that marketplace works and the other did it. So we actually pivoted to the business side of that marketplace and turned that into a better b software and then now ideally and so the constant thread through there is just kind of looking at problems and then figure out solutions, and that's sort of what keeps driving me. And it was ten years ago.

Travel bookings was a really clunky sort of thing. Really wasn't. It still wasn't as sophisticated as it is now, and it's still got a long way to go, to be honest, but. That was a problem we're trying to solve.

At one percent of that market had an online presence with live availability. So we were like, that's a big problem to solve. So you had Yonder, you built it, you scaled it, and you exited to Thomas Technology, the big sort of travel software company in the US. What did that sort of experience teach you?

Very concentrated period of time that that all took place, and it must have been a real whirlwind. Well with covid in the middle, which for a travel related business, really helped us inspect what we're doing. What I learned, I learned like what a good market looks like, what a good sales motion looks like. I think one of the things we struggled with was some of the economics of selling into the US from New Zealand for this sort of type of price point that that product was, which is part of the exit or our exisition by to this othern American based technology companies.

They already had a sales force so that was just in the US, and we had better technology and all the big cold of customers in the side. So that sort of solved that problem. But so so and yes, a problem sales motion. And also there was a really long tail in a way laggards in that industry, and I realized just personally, I enjoy really working with quite just a bigger set of people looking for an edge and that sort of adoption curve.

There's like a bigger middle of people who are really looking for that edge just because it changes a lot of the go to market motion it terms, our customers are constantly pushing us to be better, which is awesome, like that customer feedback to help shape what the product and. Value that we can deliver. So I learned all of that and actually from that sort of almost did a personal reflection of Okay, I really enjoyed the space. I want to work on something even bigger next time, and worked out the kind of series of things that I would be sort of grading a new opportunity on and went through a few different ideas and then this ideally idea, and then the market research space just sort of tacked a lot of those boxes.

You know, I had track Suit on the podcast last year, and there's a lot of sort of parallels with that company, including the sort of you know, the origin story off them, you know, some of the same people sort of integral to the beginning of it, but you know, market research, brand tracking all of that stuff. Traditionally has been a very expensive industry. To get market research done on an area you want to target, or to get do perceptions research on your brand costs of fortune, particularly for New Zealand companies, So therefore a lot of them don't do it, or they go into a very small sample size and they don't get really good quality results and sometimes they're launching new products on the back of that, so it can go very wrong.

So I guess that was the opportunity. You saw again, a huge market and people within that market who want to get an edge, who want to pursue excellence, and you saw an opportunity to do it. Well, there's two hundred and forty billion dollars spend on market research in total. Wow world.

Now there's a wide range of market research methodologies, rating from sensories sort of research, so it's like sort of really tasting things and using feedback from people. That's one end of the spectrum all the way to like big quantitative studies. But if we look at the sort of market that we're in, the type of research that we do, it's a forty billion dollar industry which is also quite big, and ninety percent of that spend is on these big, people heavy agencies, so Canta.

Nielsen's that sort of thing. So ninety percent is still done by a very people dominated process and only ten percent is done by technology platforms like US, So. There's a really big opportunity there. And AI, I mean when AI, when it first came along, we started looking at all the different industries that it's really going to disrupt, so white collar workers.

In this case, it's gone through law and legal and health. So market research is just one of these other ones where it can start to do parts of the role that that people may have been doing. Because we started after GPT three, I think of still GPT three that our first version of ideally was based off what was that you know, two and a half years ago, three years ago, and you know, And so that's the perspective that we've gone into this, which is sort of how can you sort of automate a bunch of the process ranging from analyzing data to actually now more complicated things like even reasoning the data, Like if you take a behavioral science model, how does the things you've just learned from people stack up?

Like what matters and what doesn't? So we can start to do more complex analysis, still led by a human. But what it means is you sort of shrink the timer might take to do a research study from six weeks to overnight, and we're actually working on some other things. That means you can actually do that in minutes instead of six weeks if you've already got some of the data with you.

And what I love about that is because when you create like a whole new possibility, like technology possibility, you actually then create a whole new set of ways to do things. And I'm giving you an example of that. If you think twenty years ago you had a phone in your house, there was one landline, you had to take turns to use it. Imagine you want to talk to one of your friends or family overseas.

Well, it's going to cost a lot per minute, and so your parents in the house might be really looking at you to make sure you don't spend too much money calling. So that was twenty years ago, and now even my kid has got a phone, and it sort of shrunk the view of what overseas or a border looks like. I mean, they chat to friends overseas all the time. Well, that's kind of the shift and behavior change that we're sort of creating with this new possibility with technologies, we're turning research from something that's done occasionally into kind of always used and by a range of people.

And that's pretty exciting. Through this possibility with AI baked into it. I've seen this in action myself. I've done a bit of freelance work recently with your team at Ideally, and in this case, it was market research into Australian's perception of the banking industry.

Fast changing industry. You got neo banked, digital banks coming in, you got you know, there's hardship, there's cost of living crisis. So people need to understand in the banking industry what is driving behavior, switching behavior, that sort of thing. And your team went out, we designed some questions, a demographically representative sample I think of about four hundred and twenty people something like that, which is important.

And I was like, okay, well, i'll talk to you at the end of the week when the results come in, and literally I think first thing the next morning, it was like, here's the results, here's the analysis of the results. Here are all the graphs. Here's how you cut it any which way you want. I've just blown away by that.

Just how well that, as you say, has truncated the whole process of doing the research, assembling it, making sense of it ultimately for customers who need to make a decision quickly. Yeah, that's awesome. And you could be doing then having more questions off the back of that study, and then testing it again the next night, and so that's then a whole new possibility that was just hard to do before. You talk about a living data center.

Is that what you mean? So you've gathered this data, you don't even just discard all of that. You're constantly building on all these really useful snapshots of consumer behavior that you're gathering. Yeah, I mean the mental model of what it was like today or before this was.

You've got analysis in one PDF, and then in share points, and then another PDF and share points and a different folder, and then to make sense of the previous work you've done, you have to keep going into these different files, open it up, and then connect the dots. But we forget out of way to connect the lots within the platform, and then it sort of ladders up and grows your understanding of. Whatever audience are interested in. And your example, Peter, what were you interested in the Australian study that you did.

Well, this was we wanted to get a snapshot across you know, the population, so you know, what are young people thinking, and they're you know, I think the data showed skewed towards them. They're more amenable to switching to a neo bank because they don't necessarily want to go into a branch as much, they don't need as much support. But when they want to contact the bank, they you know, a lot of them want to talk to a human. They don't want to necessarily go through chatbots, whereas the older demographic you are more interested in still having a local branch and ATMs.

And they're like, so it was really like a let's look at the whole population and see what the trends are here. And presumably you have you basically have panels across the world because you operate around the world world. You've got panels of people who have agreed to answer these at the drop of a hat, at very short notice, answer these surveys on all number of topics. Yeah, and so I should probably clarify that part that part of like going out to people around the world.

We're following the same sort of rigorous research methodology that all the research agencies do use. We even use the same service that you just mentioned just now. It's not something that we run. We sort of pay someone else who have these relationships with people all around the world and remunerate them for doing these surveys.

It's just how we use that service, which is sort of a special source both the asking the right questions and then making sense of the data on the other side. But yeah, by tapping into this particular panel, our clients are testing literally all around the world and in different languages, and AI is also great at helping that translation process. I think, you know, the crucial thing is knowing the right questions to ask. So there's a real art to that, isn't there There's an art to surveying and generalize.

We've seen with political polling in New Zealand for instance. You know how you reach people, the questions you ask that you know, it's actually a science underpending this doing it properly. Yeah, and so that's where we're bringing AIM to sort of help use that science to analyze the results as well as asking the right questions up front, which means we can open up this to people who aren't research experts, enabling a new set of people to like have that study that you mentioned just before, banking at the finger tips of anyone in an organization, which is pretty awesome.

It is. Yeah, that's hugely valuable. One of the distinctions you make is that you are actually using human generated data. There's a big movement in market research towards so called synthetic data.

I actually caught up when I was in London. When I caught up with the tracksuit guys, I went across town to visit a guy called James He who grew up in New Zealand. He is a company called Artificial Societies. He's taking a very different approach where he's gone completely synthetic data.

It's based on the profiles of real people, their social media profiles. But he is then assembling all of those profiles and essentially asking questions of these profiles and trying to build a sort of an artificial society and what would they do, how would they behave under certain circumstances. He wrote one of the first sort of major papers that was published in the scientific literature about that. He thinks he can get pretty good results eighty percent plus sort of accuracy.

So there is really this sort of change going on, isn't there where a lot of companies, including AI companies, are using synthetic data. You've taken a different approach. You still want the human at the center of the data gathering here. Yeah, And I think synthetic is going through maybe an identity crist because there are so many different versions, and I think there's no one right version.

It depends on the type of decision that you're making. I actually came across a guy just recently who made his own digital twin or synthetic version of himself, gave all this health information, he did personality tests, leadership tests, and sort of just fed it all its description of him that's not necessarily gleaned from like an Instagram social profile. And I can see a set of use cases for that too. And so the version that we are building, I mean, when you mentioned eighty percent accurate, all, actually that's not good enough for the type of innovation work that our customers are trying to do.

They're trying to find the nuance, they're trying to find opportunities, and that comes from really statistically representing society and it's understanding what's culturally happening today. It's about understanding people's needs and behaviors, not just sort of what we necessarily put on social media. And by understanding those and then wading through it, and then you can sort of start to size and find commercially attractive new propositions to go after, and then start to test those propositions to really understand if and how they're going to land, and then ultimately develop confidence to spend millions of dollars to make new products or new marketing campaigns and bring them to life.

And so just finding the most quality signals is really important for those types of decisions. So our approach is synthetic, is actually, with all the human information we've been gathering through the survey, we train a model and build some tight guardrails around it. The model is therefore actually a way to interrogate that data that you've already got, rather than try to predict outside of the data set. Yeah, and I think that's a really valuable approach.

You know, you're working with a lot of big brands like to Telstra, Door, Dash, Burger King, So presumably in practical terms, you got it a Burger company they want to release a new product or something like that, a new type of burger, some special that's going on the menu almost instantly, then you can ask a bunch of questions of real people who are demographically representative of the sort of people who might wander into a Burger king store. And no instantly, but not just that.

Over time, if you launch that product and then revisit it, you just you can keep adding to that all of the data that's related to that, and presumably bigger trends about fast food and what people want to eat as well. Yeah, nicely described as exactly yes. Yeah, so that's the ultimate use of it. And I think the you know, where the value is being seen here is it's no longer a fifty thousand dollars research exercise.

You know, you're truncating down the cost of doing that and therefore allowing you to do it on a much more regular basis. Yeah, more regular basis, and you can actually start to test things that may not have hit the bar of let's say, fifty thousand dollars before. Could be a lovely example, say. She here a New Zealand example with Goodman Fielder where I mean, have you been in a room before where everyone's got different ideas, and you're like, well, how are we going to decide who the ideas is good?

And who ideas whose idea? Are we just going to part and lot to do? Like this can be a bit of a war. But if you're can actually start to bring evidence into that conversation, then you have a much higher quality conversation about how to move things forward.

So that's what happened in this just one example actually all of our customers, but one example that comes to mind is it was a break new breakfast to drink by Goodmanfielder just for kids, and so one of the younger marketers in the team was like, I know how we can market this. We should get Steven Adams, a basketballer, and Stephen Adams loves anime, and so to convince sort of a wider marketing team and don't necessarily even though who Steven Adams and is an anime is also quite a wild idea, they were able to get back into this idea and then it's actually won a bunch of opening awards.

It's been really successful, and it's partly because I've just been different, but it's also deeply understanding who you're actually trying to sell to what's going to cut through and resonate, and so that empowered We'll just change the balance of powers from the team to ultimately get the surface the best idea, not just the artists voice. I could see that really being useful in like a newsroom where you know where I work, where everyone's got really strong ideas about what should be the focus of the front page.

But whether there's any evidence behind it, whether anyone wants to read it as a different story. I guess one of the criticisms of artificial intelligence and generative AI in particular is it it's very good at its summarizing and making sense of what has come before. And I guess when you apply that to market research, is there a risk that the companies that are using this on an ongoing basis become a little bit more conservative, that they're less willing to try radical new ideas that can create game changing new products, that they just go okay, continuously, we have a live feed basically into the brains of our potential customers.

Let's not do anything that they don't want. That's a real problem, and it's a problem that we have therefore chosen on our approach to use responses from real people around the world and then make sense of it with their AI. And it's because it removes the issue that you're describing by putting your own human ideas that you've come up with and putting it out to people and seeing what they think, and then using our idea AI to sort of make sense of the results you get back. That's kind of the sweet spot of the mixture of the humanness as well as a AI to make sense of it.

And which is why when we talk about synthetic that the use cases are not all that situation. I came across a video fell by someone, well, actually a number of stories where people have tried to test different ideas with like an LM created synthetic personas, which is different to like something trained on a human dover set and everything just get force into the same thing. And I saw this video where it was like, tell me a car brand, tell me another car brand, and you asked all the different LM models and they gave you the same sequence of things.

Because it's all based on the predictiveness of the volume of words on the Internet, which is then driving the predictive. Algorithm to give you the most likely outcome. That's so what you've described is absolutely relevant, which is why you need sort of a different approach to find those novel ideas that are going to cut through in like today's world and got the biggest success which aanced the success when you put them out. There, which, as you say, is really the secret source of dearly and what you're doing with your own model, how you're tweaking it and developing it for those purposes.

And yeah, I guess you've had really strong growth, particularly in the US market and market teaming with brands, very competitive sub markets there that people want to understand. So ideally canvas, I guess is the core product now that's serving customers over there, maybe talk us through Is that what I was using likely for the console to make sense of all of this research? Is that's what the is that at the height of canvas, it. Was adjacent to campus.

I think of canvas as being a collective understanding of the category that you're interested in. It in your case, it was like the banking category, and everything that you do sort of adds up within it. But the core of the canvas is an understanding of your consumer segments. So those different cohorts of people ranging from like gen Z savers in the banking The banking example is share before to maybe boomer spenders.

I don't know. However, you think about consumer segments, and then there are other dimensions that are really interesting, like occasions or category entry points, sort of different marketing language of thinking about when and where people are, like consuming a bear, for instance, like Corona is a very much of a summer hot day bear. So that's the occasional on the moment that they went on, but there's of course many other moments that you drink beer. So you start to build a sort of fundamental parts of category, and then we can start to understand where the competitors are on that map.

So when you come to try and find new opportunities, you've got an idea of where the demand is, who's playing in it, and then ultimately try and find some opportunities for your brand to play in, whether it be a marketing campaign that appeals to certainly might pain point. Or a new product. You're obviously disrupting a big market you said forty billion dollars. You know, a segment that you're playing in.

What's been a response from the can tours of this world. The big market researchers are they are they innovating, scrambling to change how they do things to be much more efficient and responsive. Yeah, scrambling as all relative if you've been in a large enterprise businesses like you described versus a more nimbles startup like us, So yeah, that'd be scrambling somewhat. Cultrics I have made a lot more noises than others on the new direction that they're trying to go some similar things to what we're also doing.

They're talking about doing in the future, but there's actually a different belief here, and the belief is where the balance of power is going to be in terms of who's going to be using these tools because a lot of their models they're not technology first models. They are trying to use technology to make their people providing the service more efficient and therefore productive. We're trying to flip that on its head and actually put the tool into the organization. So it's the.

People in the brand, the marketers, the insights people, the products people doing the work directly, and we feel our thesis is that's actually going to get better outcomes because the more people using it in their day to day work often is going to get better outcomes for that business, which is a general theme that we see from our customers and how they think about using customer insights in their day to day work. I think that's really the most powerful aspect of it. I've been in organizations where you get called to a meeting and they present all of this research that has basically been done to you.

You know, you don't know what questions were asked, you didn't have really any input into it. And then this external company comes in and presents the findings about a business you know incredibly well, and often you sit in there going, really, is that what you found? So I love that idea of putting the tools in the hands of the people who understand the business the best. Yeah.

And so there's a really interesting thing here where context is like everything. Context is actually going to drive getting outcomes that match what you're trying to do. I've just been at a conference in Miami. There's all sort of the cmos and marketers around the world sort of talking about the future of marketing, and they're all looking for systems that can consult workflows that are integrated and ultimately using as much kind of knowledge and information within the business as possible to be able to do those things.

So it's really interesting, well because I think we're going to start to see the rise of more sharing of information in context rather than these tools being islands on their own or the example you just shared before of like agency doing it like they would have had limited information about what your business is to kind of connect the work they're doing what you need to use information for. And there's definitely a lot of things being turned on their head at the moment in the world of AI.

And we saw in February the so called SaaS apocalypse where in a space of a week, you know, hundreds of billions were knocked off the value of software companies. A lot of that value came back. Was it slightly irrational? There's a lot of irrationality at the moment, but there is a seat of truth today.

You know, we've built our tech industry really on the back of the SaaS model software as a service zero and Timely and vend and all these great companies mainly business to business companies that charge your subscription for a piece of software and that does something really smart and I guess you know, ideally as an example of one. I'm sure it's a new piece of software, but it's sort of potentially undermining existing lucrative business models and software products are already out there.

That's our goal, and so a couple of different ways we can do that. One is actually how we charge or price our offering. Traditionally people charge per seat like a license fee, based on a number of people using the product, but we actually charge based on how much we use, and so it really changes the nature of like value. So that's one big disruption that's coming, which I think is good for everyone.

It's good for customers they just pay for what they actually get value for, and it's good for us because we're focused on delivering more value regardless of the number of people in the organization. So that's one and the other one is valuing companies that can develop like depth and defensibility in something that that's really coming through because AI, it's sort of easier than ever to actually build product. I saw someone recently got an AI to look at another systems like documentation and API build as specifications around it, and then a couple of days that had completely rebuilt that same product.

So it's easy to kind of build functions and features. But it's harder to build proprietary data sets and become like an essential part of a workflow. So that's the stickiness factor. So yeah, it's just another evolution of how.

Businesses are valued. Really is the impact of this AI. One thing you said in a recent interview I think was you mentioned fifty new supermarket products to disappear within a year, which is just a staggering statistic. And sure, some products just for unforeseen reasons will fail, but how much waste there must be in the clothing industry and you know on Netflix all of these companies, he's putting stuff out there that that sort of fizzles.

Yeah, and fail means failed to meet the expectation by the financial expectations to sort of keep producing it and making money out of it compared to investing at money somewhere else. And that's the new products. And then for the marketing version of that, as I mean, how many boring banking and adverts are there, You start to sort of gloss over them, and banks are sort of notoriously difficult for them to take a bit of risk. And so if you can just be more informed in doing that and actually solving a genuine customer need in the process, I think, yeah, that's better for everyone.

So when you see sort of ideally in five years, James, is this a sort of a real platform play or is the real value here is an enterprise data asset? Are you an acquisition target by cantor or one of those sorts of players. Where do you see the trajectory going in the next few years. All three of those we need to be achieving simultaneously.

So genuinely tho three things that you mentioned very much of our roadmap. Like if we build something great, but it's easy for that customer to shift somewhere else, then that's not necessarily good for us as a business. We want to sort of keep relationships with the customers and then keep retaining customers. And so the data that you just mentioned defensibility is really important building for other strategic acquisition in the future or some other sort of exit.

It's certainly something we also have to be thinking about and how what value looks like in actually that new world, Like like, what's this new if I bring this back to the marketing world that we're in, what is that toolkit of the future marketing business? And where do we fit and how does it connect and to others, because then you can start to. Create your potential for strategic alliances. Well, good luck with the heads again.

Congrats on that decent raise. It'll give you plenty of runway now to expand as you are around the world. So good luck for the next period. And thanks for coming on the Business of Tech.

Yeah, awesome, follow you guys for a while, so it's a pleasure mind you'll also been chatting to you Petter. Thanks James. So that was James Donald's CEO of Ideally. A few things I keep coming back to from that discussion.

Number one, this year scale of the opportunity forty billion dollars in market research spending, ninety percent of it's still done. The old slow, expensive way AI is coming for that and fast. Second, the debate around synthetic data is genuinely fascinating. Different approach is being pursued here.

It's not that synthetic is wrong, it's that the use case matters enormously when the stakes are high and you need nuance, you need real people. The third, that concept of living day is one of those ideas that sounds obvious when you hear it. Why would you ever, let valuable customer insight go stale in a document or a database no one reads, but for many companies that's all that's on offer when it comes to market research. And finally, the saspocalypse angle is one I think we'll keep revisiting on the show.

The old SaaS model charge perceipt, extract maximum value is being challenged by AI native companies like Ideally that charge based on usage and value delivered. The shift and how software is priced and how companies are valued is going to reshape the New Zealand tech sector in ways we're only beginning to understand. So if you enjoyed this episode, please subscribe, leave a review, share it with someone who works in marketing insights for product development, because this one's for them.

I'm Peter Griffin. Thanks so much for listening to the Business of Tech. New episodes dropping every Thursday. I'll catch you next week.