Head counts in your team equals growth

Yash Shah (00:00)

Hello and welcome back to Building Momentum, the show where we peel back the curtain on the exciting and often chaotic world of building a successful technology business. I'm Yash, your host for this show where every episode we bring you the stories and strategies of people who've been in the trenches, conquering churns, scaling their teams and building products that people and businesses love. In this episode, we'll be chatting with Josef from Otterly.AI

Otterly.AI is an AI search monitoring platform. We're excited to hear their story and the lessons they've learned along the way. We'll be dissecting the wins, the losses and everything in between. So buckle up, grab your headphones and get into the dive in the world of building digital products. Hey Josef, how are you doing today?

Josef Trauner (00:41)

Hi, Yash. Thanks for having me today. I'm doing fantastic and I'm really excited to be here today with you and talk about large language models and search engine optimization in the future.

Yash Shah (00:53)

Yeah, no, absolutely. Thank you for joining in on Building Momentum. so let's start with what is the big bad problem that Otterly.AI is helping businesses and brands and people solve? Like what's something that should happen, but it's not happening well enough. What are you fixing?

Josef Trauner (01:04)

Yeah.

Yeah.

Yeah, I think all we are here, we know that there is this new large language model thing on the market since quite some time. And what our customers were thinking, basically, while we were also developing the product, were there is, we have no way to track how our brand, how our products

are observed or available in large language models. So because we know they are trained at a lot of data on the internet and you can ask them quite simple a question about what's the best running shoe, what's the best car and things like that and the large language model will provide your answer. So all our customers out there have no idea how they are there. Is this a positive sentiment? Is it negative? Are they not mentioned at all? So what are the things that...

Our customers, and basically that's where we started with the problem, they have this question, we have no idea if our brand is visible in Chatgpt with your perplexity AI. That were questions what we got asked from our customers.

Yash Shah (02:09)

And you're helping brands and product owners, product managers, sort of growth market is essentially solved that to an extent in the sense that Otterly.AI helps me monitor for what questions and for what topics and for what categories, for what kind of people does my product or my company, my offering, my service offering, whatever the case may be that appears and comes up. And then do you also help them sort of

Josef Trauner (02:23)

Yep, yep, yep.

Yash Shah (02:33)

give them some tips as to here are some things that you should do to be able to appear in certain other categories and so on and so forth. And would it be fair to say that what SEMrush does for vanilla SEO. I'm so sorry for that if you hate those sort of analogies, but we've got to sort of. Sure, no, please go ahead.

Josef Trauner (02:47)

No, no, I'll go, I'll go, it's a good one. I'll clarify it later. No. Yeah, no, no,

exactly. So I think where we are started is we know that this time of just having pure keywords and doing SEO is starting to end and there is something new coming what we call search prompts. So we call it very analogy between keywords and search prompts. People in the future use search prompts, in the past use keywords.

Then to your next question is, do we also provide actionability? I think this was the most painful thing what we had the last couple of months while we were developing the product is that all the large language models and all the, let's call it the iSearch engines are really a pure black box.

Yash Shah (03:28)

Mm-hmm.

Yeah, they are black boxes.

Josef Trauner (03:31)

So that's our

thing, our tips that we are developing are by research, we're just trying to find out what could be done to optimize the perspective of a brand or a product within the LLM. So yes, to your answer, we provide those tips, we work on those tips, but from a pure, what our product is currently doing is a pure monitoring tool, we just monitor where are you...

Yash Shah (03:52)

Hmm.

Josef Trauner (03:54)

Is there some change happening? yeah, with our tips, you potentially can do it, but we not have yet explicitly say yet, because I hope we will come to a point where we have a clear evidence that with this five actions, you will be ranked higher on that thing. And to your question about SEMrush, basically we have already a partnership with SEMrush where our app is listed in their app store for basically enriching their...

Yash Shah (04:08)

you

Josef Trauner (04:18)

SEO keyword product with a new AI search monitoring capability. So that's why I was laughing on your thought about, about, about SEMrush. But basically it is. So when, you were working in the past with SEO and optimizing your websites to be ranked higher within Google, that was, that was the approach in the past. And nowadays you have to look how I am ranked higher in answer or how I am ranked better in in the AI search agent.

Yash Shah (04:20)

and

Hm.

Yeah.

Yeah.

Josef Trauner (04:45)

Yeah. And I think in general, just one sentence, and I think in general, we are facing two disruptions in short term and one in mid term. The two in short term is, as I said, you're missing organic traffic, it's declining from Google because the sources are coming from different. And the AI search experience is somehow the alternative to search within Google. That's the short term, a very concrete problem our customers are facing.

Yash Shah (04:46)

Yeah, no, absolutely.

Yeah.

Josef Trauner (05:09)

And we're thinking a little bit long-term or let's call it midterm because I think there is no long-term in AI. There is always so fast changes that we can call it midterm. It's agentic user experience. So that you really no longer using any kind of search interface. You're talking to an agent. You have this agent-like experience and the agent is performing the analysis, the search, the shopping tour, whatever you need for you.

Yash Shah (05:15)

Yeah.

Yeah.

Yeah.

Josef Trauner (05:36)

And that's, think, is the

midterm. And that's why we have to be all prepared to how our products are, at a certain point, known by an agent. Because the agent will purchase for their users a product. And I want to be that product.

Yash Shah (05:43)

and

Yeah, yeah.

Agents will make some decisions also on behalf of the individual as well and we want to be that. And to the point where these large language models like ChatGPT or Gemini by Google, all of these are fairly popular and being used by people, but they are black boxes, right? Because it's sort of like trying to figure out the ingredients and recipe just by looking at the dish and then tasting it, right?

And it is a difficult thing to do, lots of experiments to run. But once you sort of cracked it as to how does this work, then it can be a tremendous opportunity. And so just for our listeners who may not know what is Otterly.AI, if you can tell us what is the status quo? Where are we today? Where have we been? By when did we start? Where have we reached up until this point of time?

Josef Trauner (06:37)

Yeah, it's a very good question because I think when we initially started with Otterly AI, it was just really a thought that this thing will come. And I think the early, early thought concept and prototypes and playing around with it was done with pure LLM. So let's go a little bit back before.

I think it was in the early 2024, so one year ago somehow. Basically, you asked ChatGPT a question, it was just answered based purely on their trained data. We can remember where we asked about recent things and it says, have a knowledge cutoff. So this term was knowledge cutoff. So for us it was clear, okay, this is something we can do. So we tried to...

Yash Shah (07:00)

Yeah.

Josef Trauner (07:19)

to find out about brands and things in there. And then we said, okay, this is somehow not really a good product, good look and feel. If you always get a knowledge cut off, for example, if I develop a new product and it's then two years ago, I cannot ask this question about my brand. So this will not work.

Yash Shah (07:26)

Hmm.

Josef Trauner (07:37)

Then there was this moment, I think it was around May last year or June last year somehow, where Google AI overview started. And what they already did, and I think Bing by Microsoft also did it quite early, they enriched their data by actual recent web searches, which is a technical term is RAG, R-A-G, Retrieval Augmented Generation. So basically what the LLM is doing, you have this large language model.

Yash Shah (07:37)

Yeah.

No.

Yeah.

Yeah.

Yeah.

Hmm.

Josef Trauner (08:01)

It finds out, OK, someone is searching about actual news. It searches the internet using a typical search engine like Google or Bing, extracting the information and puts this information into the context window. And now tries with this new information plus the large language model to answer the question. And that's what we then

Yash Shah (08:05)

this thing.

Yeah.

Hmm.

Hmm.

Josef Trauner (08:20)

So with Google AI, obviously it works way better because there are also recent events included. And then I think quite soon or at the same time already perplexity started that they had this immediately where they provide the sources and citations from the basically live internet, call it that way.

Yash Shah (08:34)

Yeah.

Yeah.

Josef Trauner (08:36)

And I think in summer also, ChatGPT then started with the search GPT, is basically then in end December was then the later into to include it into ChatGPT. coming now back to our journey. So where we started with the Otterly AI was a pure large language model analysis. And we say, okay, this is not something we can launch on the market. This will not work out because it's...

always a little bit cut off and we are 100 % sure that all the providers will implement this, that you don't have any knowledge cut off problems anymore. So that's why we did, okay, let's we have to work a little bit on the product that we are supporting all the things in the future. And that's then when we basically finally last year in autumn, somehow in autumn, had this first working

Yash Shah (09:05)

Hmm.

Josef Trauner (09:19)

prototype that was, OK, this is now of a data quality which actually can provide real value to our customers. Because this is something, this is where my nature is coming from. So I built already a SaaS company, sold it two years ago. And.

I think there I got the really important experience that you have to provide value to the customers. You can have the best research product, the research thing you have, but you have to provide actual value to your customer. And that's why I was always saying, okay, at this stage with just doing pure LLM monitoring of a brand, it does not provide actual value to our customers. That's why we iterated a lot and that was also a little bit the approach we made. We really did rapid prototyping.

Because when we started with the idea concept of Otterly, I always framed this nice sentence, I know it when I see it. So we just don't know what do we really solve and how do we solve. We have to build some small pieces glued together, start it and say, okay, this now looks into that direction we think and then iterate again over it. I think when you're touching and working with such a...

such a rapid changing concept and such a pure new concept you have to do rapid prototyping that the product evolves on the way because yeah and you're not planning for okay how we're handling 5 000 users just don't mind in this stage you really have to do rapid prototyping does the product somehow at least at least solve for my use case the problem and that so that yeah yeah yeah yeah yeah

Yash Shah (10:37)

Yeah.

Yeah, that's to work at one unit right for one unit it has to work.

Josef Trauner (10:46)

Do things that don't scale. Just work on this. I think even we started with when we did this first analysis was just a script and just reported to a spreadsheet because we say, we don't need to have a user interface. We just need to see some data and every one of us can imagine how data then looks in a nice chart instead of ugly spreadsheet chart. So that's not the problem to do this later. So I think this was a very interesting stage to really

Yash Shah (10:59)

Hmm.

Yeah. Yeah.

Yeah.

Josef Trauner (11:13)

to not dive too deep into a rabbit hole rather than staying out and touch a little bit on the fields around here. How we shape this first prototype MVP into the right direction. And on the side, having the market constantly changing with technology.

because this is what was really crazy during the journey. As I said, we had this pure large language model, then Google introduced something different, then Perplexity changed something, ChatGPT again brought something in, a lot of other models popping up, going down, and so you have to really always monitor where the technology is going with such a product development.

Yash Shah (11:29)

Yeah.

Yeah.

Yeah, absolutely. so that was actually going to be my next question, right? it seems like the technology is moving at a faster pace than the market has time to adopt it. so even like while the market is adopting and learning and the technology is also becoming better, how do you approach customer

feedbacks and customer reviews or research also, right? Because what are their needs today could be solved a month later because of some third party thing that could have happened. What their questions are today or what their behaviors are today could completely change six to eight months later.

Josef Trauner (12:27)

Mm.

Yash Shah (12:29)

and companies like yourself, So companies like yourself might be like in other categories, they might be the reason for that change. So how do you, what are the questions that you ask that that never change, right? So how do you actually approach and figure out that this is the problem statement that I want to solve for? How do you approach that?

Josef Trauner (12:46)

Yeah, I think it's there is really important to zoom out multiple times on this 10k view. And I think you really have to ask yourself, what is the problem which still exists in two, three and five years? And really completely ignoring the underlying technology change for a moment. say kill it for a moment because what we really know is that people wanna, for example, sell their services, advertise their products and things in the internet.

Yash Shah (13:01)

Hmm.

Josef Trauner (13:13)

That's what still will be the same in three to five years. So, and I think the only thing currently is changing is the way how we are accessing this information. So it's no longer directly by searching a search engine. It's gathering this information, some somehow different or even better, someone else, typical agent is then purchasing this product or accessing this information. And I think what

Yash Shah (13:13)

Yeah.

Get it.

No.

Yeah.

Thank you.

Josef Trauner (13:38)

What we here do is, that's why I also say, think we also prepared for an agentic view, is we try to monitor where in this new technology field are you with your brand. And I really tried now to say very theoretically in concepts, because in the end, what is the other side is I love to be very close to the customer because as I said already, you have to solve the actual pain point right now.

That's the completely opposite of this whole strategy thinking is that as soon as you say, we will solve this problem on this level constantly, but what can we provide actually nowadays, today to the markets to make their lives better? And that's where we then really go super, super concrete to what are currently the most used large language models and Chatgpt applications like ChatGPT by OpenAI or

Yash Shah (14:07)

Thank

Hmm.

Josef Trauner (14:33)

or perplexity, we know that gemini.google.com is used heavily, or in general Google AI overviews. So we know, okay, these are the tools. They're really used right now for gathering information, gathering products. So we have to solve a pain point right here and be prepared if the market adopts constantly that we also add tools, remove tools.

Yash Shah (14:34)

Hmm.

Josef Trauner (14:53)

adopt on the way. So our vision is to be this monitoring tool for this new era of information gathering.

Yash Shah (14:55)

Thank

Hmm.

Josef Trauner (15:01)

but

providing actual solutions today to our customers. And I think this is also the reason why we really have a super demanding market need from our customers. We get customers who do cold calling to us and say, where can we purchase licenses? And we just say, hey, we are not ready to do this on the scale. We are working. Go on our website. Try our product. So we have really this pushing market need to solve their

Yash Shah (15:19)

Yeah.

Josef Trauner (15:25)

actual problems right now, but also aware that we have to, from a strategic perspective, and also as you said, for customer feedback, customer requests, be very strategic. Does this fit into this rough direction? Because in clarity you don't know it right now. And I think you also should not...

Yash Shah (15:27)

Hmm.

Hmm.

Yeah, because you... Please go ahead.

Josef Trauner (15:46)

fool ourselves and think, okay, in three months we have this clarity. This will not come. That's what I want to say. So be aware that this blurriness of the direction will stay.

Yash Shah (15:51)

Yeah.

Yeah.

Yeah. Yeah. And blurriness is, would say that it's a price that you pay. It's the cost of doing business that exists and that the absolute forefront of technology. Right. So you are at the cutting edge of the technology and the changes and there are a lot of other benefits to that. And this is just one of the things that you as a business have to live with.

Josef Trauner (16:10)

Absolutely.

Yash Shah (16:18)

Coming to understanding a little bit about how you are achieving growth today. How are you acquiring customers? Where are you putting yourself out in front of prospects customer? Which channels have you tried? What are the channels that are working? Any lessons that you've learned over there? Any failures that our audience and listeners and viewers can learn something from?

Josef Trauner (16:39)

I think we are here in a very advantageous situation because the topic we are touching with Otterly is just really a tremendous attention-grabbing topic. what we are doing is a typical inbound marketing. We are writing some content pieces on LinkedIn. We are in webinars. So the typical simple things, but the good thing, this is our advantage right now, this topic is so pressing.

Yash Shah (16:48)

Hmm.

Josef Trauner (17:01)

everywhere we are looking around, that people are just always saying, okay, we have to try this, we want to see this there, we want to be there. And I think this is, as you also said a little bit with this advantage of being in such cutting edge area, it comes also with a lot of disadvantages, don't get me wrong. But I think this is also the good thing about, but we don't have to do any cold callings or things that we get customers. We currently really are at the breaks here and say, okay,

Yash Shah (17:06)

Yeah.

Yeah. Yeah.

Josef Trauner (17:29)

one after the other customers onboarding because I think we should not, also in terms of our product, we should gradually develop our product in the right direction. And that's why we should not push it to all the customers and artificially create more demand than we have already. So I think we are quite happy with the demand out there.

Yash Shah (17:43)

Hmm.

Yeah, absolutely. so I forgot to point out, so very, very similar journeys you and I had, right? So I also built a SaaS company, ran that for about four years, built it, raised capital for it, grew it, and then exited it and saw a complete company life cycle. One of the advantages that we had when I did that was it was a CRM for agencies and freelancers. We were able to achieve decent scale.

CRM as a space generally has a lot of winners, right? So they are not, it's not like top two or three players have like 99 % of the market share. Even Salesforce has less than 10 % of the market share, which is the number one player in the CRM that exists. And so basically the distribution of revenue amongst companies in CRM, while it is tending towards zero, but the slope is significantly

significantly lesser. However, in the space where you exist and please correct me if I'm wrong, the space where you exist, seems like there will be over the next 24 months or maybe 36 months, there be two or three winners. Exactly, let's say how SEMrush and Ahrefs are today. I don't know what to call that but are today. In the space of SEO research, there are these two companies and then

Josef Trauner (18:38)

Mm-hmm.

Yash Shah (19:02)

There's very small amount of other companies who are able to make any tangible growth. Talk to us about the pressure that it could create on you or the team that you know that if you are able to figure it out, there's an amazing outcome. And you have all the ingredients of catching lightning in a bottle. You've already done that. How do you deal with that? Is that a lot of pressure?

Josef Trauner (19:04)

Yep.

Yash Shah (19:27)

What are your thoughts around this? Because when I was running my CRM, I knew that even if I did a million dollars in annual run rate, it's a good outcome for me personally. We are a 20 people team over here in India, million dollars in annual run rate is still good. But this must be very, very different. And so talk to us about that a little.

Josef Trauner (19:48)

Yeah. I mean, the short answer is yes, it's a crazy pressure. So, and I mean, you can understand. Yeah. I mean, the long answer is, yes, we are absolutely aware of that. So I think when, when we initially started with the concept, said, Hey, let's build a microsaas small company, bring, drive some revenue. And, and we found out super fast. holy shit. This is not the microsaas. This is, this is kinda.

Yash Shah (20:09)

Hmm.

Josef Trauner (20:10)

I don't want to say that way the winner takes it all, but

Yash Shah (20:10)

Yeah. No. Yeah.

Josef Trauner (20:15)

it's a sword of this game because as you said this is a space, I mean we know we are absolutely onto something. I believe we are really under the top three.

Yash Shah (20:21)

Yeah, this is very clearly

like a hundred million dollar annual run rate opportunity. It is very, very clearly. It seems like that. Yeah, absolutely.

Josef Trauner (20:27)

Absolutely,

absolutely. And I would love to talk to someone who says, this is not a hundred million dollar, but everyone instead told me this is a more than a hundred million dollar business and said, shit, we are onto something. We really have to do something. And this, as you said, creates a huge pressure on us to also deliver on that thing because also we know, okay, this is a really, really great opportunity we have here. yeah, it's a lot of pressure, but I...

Yash Shah (20:35)

Yeah.

Yeah.

Yeah.

Josef Trauner (20:54)

I really think the good thing is we all three, me, Thomas, CEO and Klaus, I founded the last company with him. We have this experience with creating a company and at least we have this experience with the basic problems of building a company from scratch to sell it at some point. We have done this already. So I think this is a good foundation to not be stressed additionally with the...

Yash Shah (21:05)

Yeah.

Okay.

Yeah.

Josef Trauner (21:20)

Typical things you do when you build your business first time, so you do a lot of things wrong and I still believe you're doing now a lot of things wrong, but it's a huge portion less because I At the stage where are we right now without the leave my former company? I think it took me two years to be there. So that's that's that's that's really the good thing But yeah, nevertheless, I mean it's it's stressful because I think it's it's a absolute huge opportunity where we are right now

Yash Shah (21:25)

you

Yeah.

More.

Yeah.

Josef Trauner (21:47)

And I mean, we just have to run the hill upwards.

Yash Shah (21:50)

Got it. And before I let you go, want to this a little more as well. So one of the things that you mentioned was rapid prototyping and it is, if this were your first company, I wouldn't ask about it. But you've done this a few times and you read this in a lot of books and you hear this in a lot of YouTube videos that you should, you know.

build out a lean team, should prototype and iterate often and stuff like that. Can you explore this a little more in terms of when you say that you are doing rapid prototyping, how do your sprints look like? How does your team structure look like? How do your deliveries look like for people who are doing it for the first time? What does it mean? at what point can I say that, okay, hey, I'm also trying to do rapid prototyping or I'm

Josef Trauner (22:29)

Mm-hmm.

Yash Shah (22:36)

moderately successful at doing rapid prototyping. So if you can lay out this from a product and team and delivery standpoint, how does that look like? That will be helpful.

Josef Trauner (22:42)

Yeah.

I think I have a few crucial things where I believe this is important in rapid Prototypic versus small team. You just have to really have a small team where you don't have to sync with too many people because syncing means just, it's just overhead and it's just intense. think this is the first thing. Then remove, that may sounds counterintuitive, but remove as much structure as you have.

Yash Shah (23:00)

Yeah.

Josef Trauner (23:07)

because it's thinking more in time boxes. Because I think this is the only structure we all can align on, but don't set up too many formal processes like, okay, we do this in our daily standup I really removed all this clutter, which can be added later in a more formal processes perspective. But I really love to just do things in a time box manner and say, let's see what we achieve today and then sync again and really...

Yash Shah (23:07)

Okay.

Josef Trauner (23:34)

touch fast forward and really say, we implement this one thing today and not over the next sprint. I never said over the next sprint, we're not doing any sprints because sprints creates a large time box where you can fill either a lot of things in or there's a lot of problems in. Rather very often zooming out, where are we right now? Is this the right direction? Also from a zoomed out perspective, can we deliver on that? Can we deliver value to a customer fast? So I think...

Yash Shah (23:45)

Hmm.

Hmm.

Josef Trauner (24:01)

What you have to do is, I think this is the good thing about experienced founders here, you throw a lot of very good concepts from larger companies overboard.

and you go back to having a super small DIMM, working super fast paced, ignore security things, ignore things like that. It's okay, okay, this we can standardize that later. We use standardized components for that specific part later. But I think in this stage, you just have to be really fast as possible. And for me also fast doesn't mean that we have to work 24 seven.

Yash Shah (24:13)

Okay.

Yeah.

Josef Trauner (24:35)

but it just means more, we use our time really wisely and not say, let's solve this in the next sprint. I never said this within Otterly yet. We're not solving something in the next sprint. Either we solve it now or we don't solve it at all. And maybe the problem pushes up to us next week again and we say, okay, now we solve it, it's really a solve every problem. But you have to be really, really precise that you don't...

Yash Shah (24:35)

Hmm.

Got it.

Josef Trauner (24:57)

jump into a rabbit hole and then you're developing for a month or for three months and because we have to be very careful at some point and that maybe doesn't matter in the long term. So I think this is really the crucial part is to throw over the board a lot of very good structured processes and just work on this topic in an unstructured but fast-paced way. Time is the only thing where we can compare each other. So 24 hours for you are the same like for me and I think this is what you can do and say okay we can deliver on that.

Yash Shah (25:21)

Yeah. Yeah.

So like a small tightly knit ownership sort of driven group of people who don't need structures to be able to perform and then consistently sort of deliver. So the first point that when you said small team to be more productive, like I had this painful realization.

Josef Trauner (25:34)

Yeah. Yeah. Yeah.

Yash Shah (25:45)

that a team of like when we started when it was my first company and when we started we thought that head counts equals growth rate. So head counts in your team equals growth and yeah and then I very painfully realized that seven people can be less productive than four people and that was a very painful realization but thank you Josef for joining on this conversation. It was a pleasure to have this conversation.

Josef Trauner (25:54)

Everyone did that mistake.

Yeah, exactly.

Yash Shah (26:12)

with you very insightful. I'm sure for all our listeners on Spotify or Amazon Music or YouTube, please check out Otterly.ai and sign up if you are looking to perform or do AI search monitoring. If you are not looking to do it now, you'll be looking to do it in six months time. So that's okay. So we'll leave that choice up to you. But thank you. Thank you, Josef again. Yeah, no, absolutely.

Josef Trauner (26:33)

Absolutely. Thanks Yash for having me. Thanks,

Josh.

Yash Shah (26:38)

It was a pleasure and we'll see you again soon. Bye.

Josef Trauner (26:42)

Bye, thank you.

‍