SaaS Stories
SaaS Stories is my not-so-secret quest to learn what it truly takes to succeed in the world of SaaS—and I’m inviting you along for the ride! I have the pleasure of sitting down with brilliant minds and industry trailblazers to explore their journeys, uncovering the secrets behind their growth, the gaps they spotted in the market, and what really drives them.
It’s not all smooth sailing—there are challenges, unexpected turns, and moments of reflection where they share what they’d love to change about their journey. Think of it as a candid, insider’s look into the world of SaaS, with just the right amount of curiosity, empathy, and wit.
Join me as I dive deep, selfishly soak up all the insights, and hopefully share a little inspiration with you along the way—one SaaS story at a time.
SaaS Stories
Data That Doesn't Lie: Brand Tracking for the AI Era
What happens when you combine traditional market research expertise with cutting-edge technology? Stephanie Clapham, Director of Research at Latana, joins us to reveal how next-generation brand tracking is transforming how SaaS companies measure their brand health.
Despite companies investing billions in brand marketing annually, most struggle to accurately measure whether these investments make a difference. Traditional brand tracking methods often deliver unreliable data at premium prices, creating a significant trust gap. Stephanie explains how Latana's innovative approach leverages ad-based sampling, Bayesian statistics, and machine learning to provide more accurate insights into brand perception.
We explore common misconceptions marketers have about brand tracking, from the "everything but the kitchen sink" approach that leads to unwieldy surveys, to an unhelpful fixation on sample size rather than data reliability. Stephanie emphasises that even early-stage SaaS companies with limited budgets should measure brand perception, starting with foundational metrics like awareness and perception to benchmark growth over time.
The conversation delves into how artificial intelligence is revolutionising brand tracking through predictive modeling, improved data quality assurance, and automated insight generation. Looking ahead, Stephanie identifies emerging trends including growing concerns about fraudulent responses (with up to 40% of market research potentially being fraudulent), decreasing representativity in traditional panels, and increasing demand for instant insights.
For marketers tired of flying blind with brand investments or struggling with unreliable data, this episode offers practical guidance on measuring what matters. Listen now to discover how technological innovation is making brand tracking more accessible, reliable, and actionable for SaaS companies of all sizes.
What are some metrics that marketers can look at to know if they're doing it right?
Speaker 2:Getting those foundational elements, your audience and your core KPIs is really essential.
Speaker 1:What was broken about the traditional model of brand tracking?
Speaker 2:Still no easy way to measure whether an investment in brand marketing is making a difference.
Speaker 1:SaaS founders and marketers not have huge budgets for marketing or branding. What can you?
Speaker 2:tell them Any investment in brand is worth measuring.
Speaker 1:Welcome everybody to another episode of SaaS Stories Today. I'm super excited to be joined by Stephanie Clapham, director of Research at Latana, all the way from Berlin. Welcome, stephanie.
Speaker 2:Thank you for having me Really excited to be here.
Speaker 1:Thank you for having me Really excited to be here Absolutely Now.
Speaker 2:Lufthana describes itself as the next generation brand tracking. Please tell me what that means. I think one thing that we're doing super differently compared to the market is we're really focused on tech innovation. So the next generation of brand tracking. What we mean by this is that we're not resting on the same kind of traditional approaches that other companies are doing. We're really focused on pushing the boundaries and using really, you know, the the highest tech and the highest innovation possible to advance the way that we think about sampling and data collection and and kind of collecting all the key metrics for, for brand tracking to measure your health. So that's what we really mean by the next generation. It's it's not relying on, um, the things that have already been done before yeah, sounds like it's innovating in a whole new area.
Speaker 1:um, and tell me a little bit about your journey in the in the research space and the SAS space specifically. What drew you to this field? Yeah, tell me your story.
Speaker 2:Yeah, I've always been interested in, I guess, the field of kind of societal trends and human behavior. This is kind of how I started my my career. I studied sociology years ago now and that kind of led me into consumer insights and market research and I think my career really started out in a much more traditional sense in terms of traditional market research agencies. I gained a load of kind of valuable experience about the foundations of research at these, but I found that over the years my attention was really gradually drawn to more boundary pushing approaches to market research and brand research in particular. I work I worked at a few agencies doing future forecasting and trend research and then kind of moved before before moving over to kind of tech-focused brand tracking.
Speaker 2:And I think Lutana's focus, as I mentioned, is bringing together classic research expertise but with technological innovation. So this was really a natural fit for us both and back in the earlier days Lutana was already doing things that would be considered innovative, I think, even in today's standards Kind of different approaches to sampling, using advanced statistical techniques to process and analyze data and things like that. So fast forward to today. It's been really exciting working on kind of move with the times, embody new technologies into our approach as and when they were happening. I think a good example of this that we'll probably talk about later is the development of AI as well, and how we're bringing that into our product.
Speaker 1:We absolutely will talk about AI later. That never escapes the podcast. But I'm just curious. You know, it sounds like LaTana is doing things in a new and different and innovative way. But maybe what was broken about the traditional model of brand tracking?
Speaker 2:So I mean brands invest billions of dollars on marketing every year, and most of that money goes towards some form of brand marketing. However, measuring brand perception is really challenging due to the complexity and scale of the data that's required, so making it more of an art than a science. Really, there's still no easy way to um to measure whether an investment in brand marketing is making a difference, and this leads to a large share of that spending being wasted. So rather than, I think, thinking about the industry being broken, I think it's more a case of. The industry is still mostly built on quite outdated traditional brand tracking methods, and these are often really expensive and really quite inaccurate at capturing brand perception.
Speaker 2:In fact, many marketers are still facing serious issues, including unreliable data, inadequate kind of global reach and difficulty accessing their core target audiences. So Lutana is focused on delivering a very unique and innovative solution to these problems. In particular, we reach unprecedented audiences using ad-based sampling, we use Bayesian statistics to enhance accuracy and reliability of our data, and we use machine learning techniques to tackle data quality assurance, and all of these approaches result in a much more reliable, scalable solution to brand tracking. Today. That's kind of we think, surpassing those kind of, uh, we think, surpassing those, those kind of outdated traditional methods right, right, I you know as a marketer myself.
Speaker 1:I definitely agree. Branding is is very important. It is definitely one that's hard to track and measure. Um, I think I saw recent research from the linkedin um from linkedin as well as the B2B Institute that was showing brand versus sales activation and which one resulted in the most amount of income and profits, and I actually said that over the long term, brand did much better than sales activation. So I definitely see the value of it. What are maybe some metrics that marketers can look at to know if they're doing it right?
Speaker 2:I think it really depends, of course, on your brand growth level, where you're sitting at the moment in the industry, but really the foundation would be awareness, perception, consideration, preference. These kind of metrics are really core to understanding how your audience is relating to you. I think another aspect to really get right is making sure that you're tracking the right target audience view. Once you have that target audience, you can see where your growth is going, whether you're making traction with that audience over time and whether your branding efforts are paying off. So I think getting those foundational elements your audience and your, your core kpis is really essential to to starting tracking in the right way yeah, got it.
Speaker 1:You must have some pretty fascinating studies at latana. Is there like one piece of research that you're especially proud of?
Speaker 2:um, well, we're constantly kind of conducting our own internal research on research. This helps us to kind of navigate the development of our product and keep us innovative. So I'm really lucky to be at the center of many interesting experiments over the years. I think not necessarily one individual study, but I think one of the most interesting evolutions have been the switch to non-incentivized ad-based sampling. This is where we kind of display our survey questions and interactive ads that people can answer on the fly and we partner with demand side platforms enabling us to display these ads to millions of mobile users across websites and apps, much in the same way that any brand would advertise over these platforms.
Speaker 2:And this has meant that we've kind of had to completely change the way we think about designing our surveys. We need super high engagement to keep the attention of our respondents because we're not rewarding them to complete the entire survey experience. And one study we recently ran was to see if we could capture open response in this kind of opt-in environment, specifically looking at unaided awareness. So essentially we're asking people to type brand names into our survey when they're not actually being rewarded to place this kind of extra effort in. And ultimately we saw really good success in this exercise, saw really good success in this exercise, um, but with many of our research and research pieces, we also learned a hell of a lot about our respondents and about their boundaries, about what we could and couldn't do there, so this was really an interesting exercise, I think yeah, right, that must have been some really good data coming out of that.
Speaker 1:and, um, I mean, I, I know, you know a lot of organizations really struggle to get information from their clients. They usually send them really long surveys, you know, as it's coming up for renewal, so I have heard that maybe sporadically sending them. Just, you know, more engaging and less questions at a time works better. But what are you finding are some of the key ingredients that customers are looking for in a brand Like? Does trust play a big role?
Speaker 2:No, um, yeah, I mean, trust is a is a huge element, I think, not only to do what customers are looking for from the for the brand, but also where brands are looking to gain trustful data as well, on on on what they're measuring.
Speaker 2:And that's one of the core things we have with a lot of the conversations we're having with our clients, um, and potential clients is that this lack of trust is happening in the industry in general. I think not only consumers are starting to kind of lose that trust with brands in terms of the, perhaps the messaging and the positioning of brands is not quite fitting in some way. Of course, there's a lot of misinformation flying around at the moment as well, so it's difficult for anyone to trust to trust much at the moment, um, but also from brands, to trust the, the kind of solutions and providers that they're using to track their brand over time, and if the data is coming back unreliable, then that trust really erodes and and they have no real guidance on how to better position towards their audience, and so that kind of trust goes in a in a kind of sequence where it starts failing from one end and it's really continuous.
Speaker 1:Yeah, I think a lot of brands probably don't consider trust to be such an important metric, but I think it's becoming increasingly important. I've read recent research from the Edelman Trust Barometer and you're right. Like the consumer, trust is just going down every year. What are some other misconception marketers have about brand tracking and research validity?
Speaker 2:I mean, over the years we've spoken to hundreds of brands and gained a huge amount of insight into what's kind of working for them and what's not working for them with their brand trackers in particular. Um, I think I could condense this down to maybe three main misconceptions or kind of problem areas. The first one, I think, um, which we actually just spoke about, that you know the survey length getting really out of control. I think this approach is kind of the everything but the kitchen sink approach. I've seen actually many instances where brands try to measure absolutely everything in their main tracker and I think this can really damage the quality of insight that you can expect to generate there. My advice to brands is always to stick to the core KPIs that you need to measure regularly for your brand health monitoring, and then anything else can be measured separately. This is kind of why we've adapted our product to be totally modular, modularized and flexible. In this manner we don't tell brands that they can't or shouldn't measure something, but it's more about making sure that the core things are tracked regularly in a really consistent manner and then everything else can be measured as dips and not kind of regular pulses. So that's one thing that the modular approach is really really something I always try to instill, and the kind of get away from that kind of misconception of everything should be tracked all at once.
Speaker 2:I think the second thing is a fixation on sample size. We regularly have conversations with brands coming from very traditional quota sampling approaches in which they're very accustomed to talking in sample size terms and often they have this general idea that X sample size base kind of sample size will be plenty sufficient to generate reliable data over time and to do deep enough audience segmentation on their data as well. And I always find this number to be, I think, a very small for true reliability, especially when you're trying to dig down into audiences and cut the data and stuff like that using quotas and be a fairly kind of arbitrary number to land on. You know it's always kind of like who decided this was enough sample. It's often actually more of a balance of getting as much sample as costs will allow and obviously costs are always, you know, attempted to be minimal in these cases. So this is really one area we try to educate our clients on, to move away from that fixation on sample size, and I think this leads to my next point very well.
Speaker 2:The third point is that not nearly enough attention is actually placed on margins of error or confidence intervals. So, following on from the sample conversation, the only accurate way to judge the reliability of a data point is to read the confidence level of it. Um. And so having an arbitrary minimal sample size of, say, 100 people on a target audience um that is extremely important to your brand will actually provide you with very low reliability in your data, meaning that you could really misplace key brand decisions on an audience that matters to you.
Speaker 2:So the way that we kind of reframe this is that we include clear margins of error on every single data point so that we guide our clients to make confident decisions and not ambiguous ones. We also follow a margin of error led approach to targeting key audiences in our data collection as well. So rather than focusing on hitting a generic number of people every data collection, we hit as many people as we need to ensure the reliability of the data point. I think that's a really, really different approach. Actually, I've not seen that before in the industry.
Speaker 1:Right, yeah, and that's really interesting For SaaS founders and marketers out there that are kind of maybe not have huge budgets for marketing or branding. What can you tell them to convince them why branding is super important in the early days of scaling?
Speaker 2:Well, I think any investment in brand is worth measuring. So if you're starting out, you're obviously investing something in your brand. You're trying to position yourself in a new market. You're trying to uncover key audiences. You're trying to make you know, have resonation with that key audience. So any investment in your brand is definitely worth measuring whether it's working or not. Otherwise, you're kind of going in blind and not knowing where to best place your efforts.
Speaker 2:So keeping a regular pulse on kind of the basics, like awareness and perception, are really the foundational steps to measuring growth and doing so on your closest competitors as well as how you would establish a benchmark in the industry and monitor how this evolves over time. So I think two key points are to understand the industry makes you competitive and understanding your audience makes you competitive. So for early stage companies, I think it's really important to get the foundation right in terms of this tracking and then you can build from there. We always kind of use this approach with our clients. We have a lot of flexibility in what we offer, but we do also have packages that tailor to this kind of growth stage. So our essential package is tailored for this purpose exactly to measure early growth and to kind of build from there over time. You can add further KPIs. You can refine your audience later on time you can add further kpis. You can refine your audience. Later on you can add competitors. But really to start with the foundations, yeah, absolutely.
Speaker 1:I think the foundation is crucial before building any more layers on top of it, or it'll just not work. Um, I think also in the sass world you know, a lot of organizations are mainly focused on acquisition new clients. How do we scale and grow? But actually brand tracking plays a role in long term retention and loyalty as well. Is that right?
Speaker 2:tracking your progress amongst your target audience, you can easily drop the ball and fall short of competitors over time or even miss out on kind of key audience shifts that might be happening in the wider society that might be vital to your own brand growth. So I think long-term retention and loyalty require regular brand efforts and it's the measurement of these efforts that will ensure that your decisions are effective. Um. So continuously measuring core metrics like perception, consideration and preference within each of your markets and among your your target audience are key to measuring loyalty over time. Um, and really benchmarking yourself and positioning yourself to stay ahead of the industry and your competitors. And, of course, loyalty goes hand in hand with understanding how your target audience think.
Speaker 2:So at Latana, we focus on really deep segmentation abilities. This is where our Bayesian regression statistics come into play to enhance the ability to cut the data. And what we kind of mean by this is that we ensure brands can combine many different characteristics into their segmentation creation. And look at this across any lower funnel KPI with still very high reliability. In normal quota sampling approaches, cutting the data this much would always result in very small sample sizes and inevitably low reliability. So being able to measure audience in this kind of a granular way is really crucial to understanding your brand loyalty over time. That's one area where we're very focused on.
Speaker 1:Yeah, I think that's pretty good, especially in the world of SaaS where, I've heard, customer success is actually where 90% of the revenue is, so very important to be able to measure that. What advice would you? Would you give companies that are just starting out to track brand perception and probably aren't sure where to begin?
Speaker 2:I think, um, as mentioned before, I think it's always to start with the, the kind of fundamentals, and build from there. So start with measuring your awareness, perception, maybe even consideration, and then to grow your tracker from there, um, and then the other thing is to really consider what your, what audience you want to measure. So not capturing the right audience in your tracker will make it very difficult to understand if your brand efforts are actually hitting the mark or not. Um, personally, I think to take some pressure off, it's always better to choose a partner to help you, guide, um, to help kind of guide you through this process too, rather than attempting to use diy solutions.
Speaker 2:I think many brands who are just starting out tend to veer towards these solutions as kind of a cheap and non-committal way to track their brand. But ultimately this leaves you with all of the pressure to get the setup, survey, design, sampling framework and everything in between right, and getting this wrong actually means that you've wasted your time and money generating results that are kind of meaningless or unreliable. And again, at Latana, we try to make it very simple to begin tracking your brand. We guide you through setting up your ideal tracker, from what metrics to track to how to best capture your audience and then providing that totally end to end service so that you just have to wait to receive your data each month and and even have helped with you know kind of in interpreting that data and get pulling the right insights out for yourself. So really taking the pressure off is actually a kind of foundation to getting everything right and tracking correctly from the start.
Speaker 1:Yeah, that's right. I promised you we'd come back to AI, so maybe now's the time. Tell me how does Latana use AI. But then also, how do you use it in your day to day and how are you finding it?
Speaker 2:So we use AI in a variety of different ways really at LaTana. I think the first one to talk about is maybe our predictive modeling. It's kind of the core of our product, I think, and it solves a lot of the issues that traditional sampling presents. This really would foster data quality, accuracy and high precision audience segmentation. So, to go into it in a bit more detail, we built a Bayesian statistical model for this purpose, leveraging multi-level regression and post-stratification. It's a very, very complex word, but to shorten it it's called MRP. It's a regression modeling technique and it allows for more precise weighting and processing of the data. In traditional quota sampling methods, every kind of each wave of data is considered separately and each audience segmentation is considered unrelated. So in latana's mrp models, on the other hand, we recognize that results tend to be similar in consecutive waves and that the segments are not independent of one another, and this allows us to build um a more coherent narrative of how brand data over evolves over time, and this results in more kind of realistic estimates that don't jump around wildly from wave to wave and MOEs, that are margins of error that are therefore about 90% smaller than a typical quota sampling data sets.
Speaker 2:So this is one way we really use machine learning and AI to enhance our data sets. The other kind of, I guess, way that we're using it at the moment is in data quality assurance. So we're measuring respondent reliability using a wide range of in-survey behavior-based metrics and then condensing this into a single score using advanced machine learning techniques. So those respondents scoring above a certain threshold are then removed from the final data set. And I think this approach really is, um much kind of better than than the kind of binary method, which means that you're using one or two metrics to clean out respondents, automatically terminating them from the survey. For example, it avoids us kind of over-cleaning and missing genuine respondents or under-cleaning and having fraud or bots enter our surveys.
Speaker 2:So this is a real advancement. I think, on the way that we look at quality assurance. There's also the advancement of how we're automating insight generation as well. So this is more on the customer side, the client side, of how they get the data back from us. We're currently working on using large language models to build in-tool data retrieval, kind of like a chat GPT for your data set, and also an automated insight generation deck, so you'd get slides that are automatically pulling out all the relevant insights for you. Both of these tools will kind of allow us to give immediate and relevant insights to our customers without them having to spend hours digging through spreadsheets or slides to find the information they need. So I think AI can really be beneficial to the market research industry if used in the right way.
Speaker 1:Lily, I think AI has already taken so many organizations so far and I mean I kind of feel weird asking this question because we just don't know where AI is going to be in the next three to five years. But how do you envision brand tracking space evolving in the next three to five years?
Speaker 2:I think it's an interesting one, but I think there's a few core trends that I can already kind of see happening. I think number one it's kind of a negative side of it, but I think there'll be greater fraud concerns, really because as AI models continue to evolve to mimic human behavior, they really are getting more evolved in bypassing these quality assurance measures.
Speaker 2:There's a stat recently that we've uncovered that shows that up to 40% of market research responses are fraudulent. So really, this number is getting higher and higher every year, and the sophistication of these tools the AI tools to mimic this behaviour and cheat the system to get the incentives from the surveys this is really something that's, I think, a core concern to be mindful of. I think another kind of trend that we're seeing right now is there's going to be less representativity in panels, really, um so panels that the traditional approach of recruiting respondents for surveys, where individuals sign up voluntarily for panels and then and then kind of get invitations to surveys to get rewarded from, um, but these are really, um, quite outdated and the younger generations are not really accustomed to being sent these kind of very formal email invites to surveys. So I think one trend to be mindful of is catching people where they are, and this is what we do exactly with our ad based sampling. We're really meeting the respondent in their native environment, whether they're or they're on their social media or shopping.
Speaker 2:Reaching these respondents in a natural environment means that we're more likely to get authentic and genuine responses from these, these respondents, and they're more likely to be more engaged in the surveys, um. So I think that's really a core trend that's going to catch up with us, even more so as the younger generations are the core for kind of sampling group that we're looking for. I think there'll also be a greater demand in the industry for instant insight. So you know, waiting for slow field work, time turnarounds, hefty spreadsheets and PDFs, like I mentioned, all of this kind of lag in getting the data as and when it's happening, is going to be more of a concern over time. So the need for fast turnarounds on data collection, fast turnaround on insight generation, all of this AI can absolutely help in this area as well.
Speaker 1:Sounds like some amazing trends and I think you know, know you made some interesting insights there on how to target the younger, the younger populations as well. I mean, I know, you know millennials now uh make up 70 of the decision makers at a lot of corporations and uh, generation z is not too far behind, so it'll be really um. I guess it'll be really interesting how organizations change the way they they deliver their brand messages um to help capture that audience yeah, exactly yeah, stephanie.
Speaker 1:My last question for you a bit of a tradition on this podcast is if you could go back in time and give yourself one bit of advice. Could be personal, could be work related. What would that be?
Speaker 2:I think the biggest thing I've learned throughout my career, and in life in general I think, as I get older, is adaptability. So growing with a company from the early days means that you have to really dive right in, be willing to say yes even when things you've everything you've kind of learned so far pushing you to say no, um, be able to move really quickly and change goalposts when needed. Um, I have to say I did struggle with this element of the kind of uh industry at first because it felt really stressful to me. But interestingly, that's one major part of why I love this industry and especially working at Tana now is that it's so exciting. Not knowing exactly where we'll be able to go next and being a pioneer company is kind of daunting but ultimately really thrilling and rewarding.
Speaker 2:I think this actually adaptability moves into kind of other elements of my life as well. I've moved to Berlin from, from the UK just over a year ago, became a mother as well, so every element requires such adaptability. So I think, like again with with starting at an early, early kind of stage with the company, it's exactly the same really. You have to really be able to go with the flow and and um say yes and dive right into everything, which is really exciting, I think yeah, I love that advice.
Speaker 1:I think I've heard recently from people talking about, you know, with ai exploding and, and you know, being able to perform so much, what are some of the skills that humans are going to need? And adaptability was was one that came up because there's going to be so much change and it's going to hit us in such a fast pace that I think the best thing we can do is just adapt as much as we can.
Speaker 1:Exactly, yeah, thank you so much for all the insights. I think I've learned a lot about brand and brand tracking already. There's definitely a few things I'd love to go and apply. Thank you so much for being on the show.
Speaker 2:Thank you so much for having me. It was great. Thanks so much.
