How To Turn Site Search Into A High-Performing Revenue Engine — Paulius Nagys | Why Complex Catalogs Need AI, Why Patience Gaps Kill Sales, How AI Matches Search Intent, Why Search Bars Provide Feedback, What Vector Search Actually Does (#463)

Please support the show if you liked today's episode:

• If you love the podcast, please get someone else to listen, too!
• If you enjoyed this episode of the Ecommerce Coffee Break podcast, please head over to Apple Podcasts, leave a rating, write a review, and subscribe.
• Share the podcast with your family, friends, and co-workers.
• Tag the podcast on Instagram @ecommercecoffeebreak and let me know what you like about it.
• If you like the content and would like to support the podcast, you can buy me a coffee here.
• Become a guest on the show or sponsor an episode.

00;00;00;03 - 00;00;28;19
Unknown
So it's a silent revenue killer. And yes, we need to index keywords, but we need also to bridge that gap when standing intent. So if we are trying to narrow this patient gap, the revenue start to increase. Hello, welcome to another episode of the E-commerce Coffee Break podcast. E-commerce has never been more competitive. Traffic is expensive, attention is short and most stores are leaking revenue in places they don't even look.

00;00;28;22 - 00;00;48;21
Unknown
The biggest leak in your store probably site search for most brands such as Treat it like a basic feature, something that just works. But when it feels shoppers bounce cart stay empty and your marketing budget goes to waste. Today we unpack why most sites such as fail what modern shoppers actually look for in whole prints can unlock revenue that's already on your site.

00;00;48;23 - 00;01;09;05
Unknown
To break this down, I'm joined by Paulius Nagys, he is the co-founder of LupaSearch and I first search and product discoverability platform that's built to drive e-commerce conversions. He spent nearly two decades bootstrapping tech ventures and going head to head with global search giants in proving one thing real return on investment on real store data policy.

00;01;09;08 - 00;01;38;16
Unknown
Welcome to the show. Thank you. Thank you for inviting. Yeah, let's jump into it. Why are shoppers today so quick to bounce after just one or two bad searches? We analyzed a lot of data recent years and we realized that we need to define this behavior and we are calling it patient. The problem which we are trying to solve right now, according to Forrester, they say 80% of online shoppers abandon search when they can't find products.

00;01;38;19 - 00;02;03;09
Unknown
We found that customer gives only 2 to 5 chances before they bounce to competitor. And we are calling it patient gap. So it's a silent revenue killer. And yes, we need to index keywords, but we need also to bridge that gap with understanding intent. So after that, if we are trying to narrow, narrow this gap, the revenue starts to increase.

00;02;03;11 - 00;02;32;15
Unknown
Why does the classic keyboard based search tool fall apart when it comes to today's search experience and what people are actually looking for? I would say that it's days. It's like a commodity, a keyword search. It's a commodity, but our expectations are increasing. And I would say maybe it's a good example. Google and the Open openai companies, they are creating some kind of habits.

00;02;32;17 - 00;02;58;16
Unknown
So Google introduced search, I don't know, 20 years ago, more than 20 years ago, and all the world created some kind of habit. You know how we browse and how we look for information, how we find information. And 2022, it was a breaking point when people started to realize that we can create longer queries. They can create question and receive answer.

00;02;58;19 - 00;03;28;14
Unknown
And it it builds some kind of habit for everyone. And now we expect that all e-commerce search would behave the same because like a Google started to create some kind of habits and then it was like an extension to e-commerce website search. Now openness, creating new habits for longer queries. And we see that those longer queries increasing and that it creates frustration for all the visitors.

00;03;28;14 - 00;03;59;18
Unknown
If they can't find what they're looking for, they just, you know, coming and starting to discover what is around to them. And during this discovery process, things are starting to be very complicated. Mm hmm. I can totally understand that. You mentioned that the search behavior has changed, and I think everyone can relate to that. Now, obviously, somebody who comes to your store or to your site and search is a very valuable customer because they already have in mind they're probably already come with a bias and tend to because they're searching and not just browsing.

00;03;59;18 - 00;04;28;17
Unknown
So I think there are just next level and closer to the credit cards than somebody was just browsing. When we talk about search abandonment and you mentioned a number before, how big is the issue for a typical, let's say, Shopify store? I would give another stats. The Beamer Institute, Bemidji Institute is very well known institute the EU released last year is that 70% of eCommerce search tools fail on a simple query.

00;04;28;19 - 00;05;01;00
Unknown
Let me give you a real life example. Imagine the customer walks into a store looking for a couch, but your database only has sofas. And in physical world, a salesperson would say Right this way. In Digital World Search, Bar says no result and the customer leaves instantly. So before a big companies like Amazon had a huge team, literally floors of people just manually writing synonyms, synonyms, lists for every language counts.

00;05;01;07 - 00;05;27;00
Unknown
Couch equals sofa trousers, equal pants. It was expensive and slow and now new way. Is it that, you know, we can connect to large language models? We extract these synonyms automatically and we don't just match words, we match meaning, and it closes that passion gap. Now, AI is changing a lot there and was your tool Looper Search. You're in the forefront when it comes to Star Search.

00;05;27;02 - 00;05;54;27
Unknown
Tell me about the approach that Looper Search has and what kind of benefits it has for a store visitor. We are trying to not to compare features to features that are so, so much of software, this martech vertical with tens of thousands of tools. It started be a bit of insane because merchants, they are very frustrated. They need to pick the right tool for their needs or their business needs.

00;05;54;29 - 00;06;30;12
Unknown
And usually you need to compare hundreds of features, and it's static to be a very time expensive. It's a high frustration. And the we sell to that signal that the merchants have restricted to pick the right tools and we approached customers that let us not to compare features with features. Let's compare results because the result is, you know, then the goal actually and then, you know, you can easily, you know, to actually Internet is really good in measuring all the data points.

00;06;30;14 - 00;07;03;24
Unknown
So it's really easy to measure which tool generates more revenue. So our approach started to be like, we are coming, not the comparison features with features, but we are coming with a commitment that we can analyze the store and we can predict how much we can increase conversion rate via search. And that's then goal. And we realized that it's a release for for merchants because everything with the you're about is the result.

00;07;03;24 - 00;07;24;09
Unknown
Actually, that's very true. That's very true. At the end of the day, the technical administration, the tools in the background are all there to increase your conversions and your revenue at the end of the day. Talk me through how Looper search works from the side of a store visitor for our business that obviously can't see it in front of them right now.

00;07;24;11 - 00;07;54;20
Unknown
We call it Western Research, so we don't look at letters. We'll look at the mathematical distance between concepts. So even if the user types black leather jacket snatched together, we know exactly what they mean. So it's like a graph database and we are comparing distances between words. So that's that's the technical explanation visually from user experience. You know, we have a dashboard and you have a lot of analytical data.

00;07;54;23 - 00;08;24;13
Unknown
So some of the customers, they realize that search bar, it's the biggest feedback channel to connect with the visitors. So just imagine that there are millions of queries and we are collecting all this information to our dashboard and these analytical data. What is happening with the with the user behavior. And you can you can predict a lot of, you know, data driven decisions, what needs to be next.

00;08;24;13 - 00;08;51;28
Unknown
You know, they are browsing for some assortment, which you don't have. So maybe you can add to this assortment. Maybe the browsing. What is your working hours if you are omnichannel business, if you have physical stores and online, we see a lot of actually stats connects omnichannel experience that people are not looking for privacy policy return policy working hours which is actually connection with physical but not with the online business.

00;08;52;04 - 00;09;44;09
Unknown
And we realized that it's the biggest feedback feedback channel which can be leveraged and analyzed. So from user experience, I would say analytical part is a big thing. It's not about, you know, how search and discovery works, but also what kind of feedback you are receiving from your visitors. So you need to have a dashboard where all this data collected and also these agents, they can actually try to analyze all this data and to give you a lot time, you know, raw report, but actually action points, what needs to be done so algorithms can work not only with their with a search bar, but also with the data which is collected into a dashboard where

00;09;44;09 - 00;10;05;27
Unknown
we talk to ideas, always a bit of a training time, a period of training involved. How long does it take to wrap all this data into processes so that the merchants can actually work with it? It depends on industry. We again, we started to think how to define this thing and we came to the word the Search trust index.

00;10;05;28 - 00;10;32;23
Unknown
So we integrate, let's say, with the fashion store, and we see that the consumption of search is minimal. Usually the customers are coming and they are don't using our search. But actually it's a trust in search because if it was not working before correctly, so you need to build some kind of trust for your visitors so that they would start to use it.

00;10;32;25 - 00;10;55;25
Unknown
And sometimes it takes half a year, you know, sometimes it takes three months that we would see that, okay, numbers are starting to increase. And they started to use that search bar that it would help for them. So this is one example about of this search trust that sometimes data takes time. It's not sometimes it's not immediate results.

00;10;55;26 - 00;11;19;25
Unknown
So you need to build trust again. And sometimes, you know, if you have a large catalog and people are using search, then you can receive immediate results because it connects to a lot of data points, immediate result, it's a conversion and it's increased sales. But what does it mean? It's actually two parts of of this is search and product discovery.

00;11;19;25 - 00;11;44;05
Unknown
Search is about, you know, keywords and intention, but product discovery is how you list products. And we connect all the data points within inventory, warehouses, profitability, margin. And so when a visitor is typing Samsung, we need to predict which products we need to show for them. It can be refrigerator, it can be TV, it can be telephone, it can be a vacuum cleaner.

00;11;44;09 - 00;12;13;02
Unknown
But if in your region is a heat wave. So maybe we need to suggest for them rulers or all the equipment which would help in this see them if in two weeks there is a World Cup football championship, so might be that visitors to look for trees. So actually even you Samsung, you can actually predict according to this season or upcoming events what is happening in your area.

00;12;13;05 - 00;12;37;18
Unknown
So we are calling it product discovery. Actually, we need to connect all the data points and commercial people at your warehouse of your business knows everything because the only the stocked with those TVs, because they already know that football championship in couple of weeks. So everything what we need to do we need we don't need you know to connect with the forecast software systems.

00;12;37;18 - 00;13;04;06
Unknown
We just need to start to connect with your commercial people who works on your warehouse and that's it. I think this is the biggest the biggest gap which you can fill and get immediate result. You gave a couple of examples. Who's your perfect customer? What kind of industry or vertical works at best was, I think, the biggest value we can bring for those businesses who has large catalog.

00;13;04;09 - 00;13;29;25
Unknown
So if you have 100 SKUs or 500 excuse, I think that we can't, you know, show the result when there is like a small catalog, when they get when the catalog is starting to increase, 2000, 20,000, 200,000, 2 million, then it started to be a very complex system and then where we come in and show it as old.

00;13;29;28 - 00;13;58;15
Unknown
So we are seeing that bigger catalog. You have bigger complexity, you have bigger value we can bring. So this is a key parts in a business that it's a catalog. Complexity filters like thousands of filters. So usually when you have this things, there are opportunities of with there. You already gave a couple of examples, but could you get share some success stories or case studies of businesses that you have worked with and you don't need to name the brand and what kind of results they saw.

00;13;58;20 - 00;14;31;02
Unknown
The biggest results was in Do-It-Yourself Business Vertical Home and Garden, because it's very sensitive to season. It's a large catalog. There are thousands of filters and usually you need to upload those dynamic filters because there are a lot of categories of products and usually it's omnichannel business. They have physical store chains and they have online. And you need to synchronize between that and between those two businesses.

00;14;31;04 - 00;15;00;19
Unknown
And actually, it's it's not we and the I'm talking about physical world and the online world of the business, but it's a synchronization how to work you make the biggest evaluation is when you are coming to physical world and you see that customer assistance. They are using your search to help for physical customers. So it's a great example, you know, how search can perform them in physical world and digital, while at the same time that was the biggest success.

00;15;00;19 - 00;15;33;28
Unknown
Bookstores also usually, you know, pharmacy. We are calling it symptom search because when you are coming to a pharmacy, you don't know the medicine which you need. You are coming with symptom, you are coming. Let's say you have a course or you have a pain in your ear. So actually you are not coming to the pharmacy and saying, hey, please give me like this exact you don't know medicine this exact.

00;15;34;00 - 00;16;08;15
Unknown
Usually you are explaining how you feel, what kind of pain it's and then you are receiving a drug. So it's a symptom. So you have some kind of symptom and you would like, you know, to find the best recommendation. So recommendation software is merging with search software. So this is also one, one great example. And the biggest wins in those businesses is where we have a symptom symptom queries.

00;16;08;17 - 00;16;36;10
Unknown
Walk me through the typical onboarding process of a new user. How does it work? It's integration, so it's a blessing curse because for us, because when that when we are starting to talk with potential clients and they are very interested and then they are saying, let's do it, and then sometimes they realize that they would need, you know, to integrate to our software to it's not like, you know, just one click and you have all the software.

00;16;36;10 - 00;17;11;03
Unknown
Usually you need to fly next to the API or plug in to connect all the data points. So it takes time a week, sometimes two. So usually it's onboarding process. But but what is interesting in this part that in the market, I'm not sure, but I think that there is no any competitor who would give free of charge without any commitment demo with a real product, you maybe for enterprises and for very, very large customers, some of their search competitors, they are doing that.

00;17;11;05 - 00;17;38;11
Unknown
But let's say you are a mid-sized business, you just would like to try it out. Does it work or no? And usually you can see only with this in tentacle data demo, what we do, we just ask just one like small file of your product field and we are showing a demo with your your product and you could see the difference and you could see that all the problems which you had with search is already solved.

00;17;38;14 - 00;18;09;01
Unknown
So you are more committed that okay, I can go and to integrate that because it would cost for you, you would be highly disappointed if you the software you signed the contract you integrate and after two or three weeks you realize that it doesn't solve your problem. So usually it would be good, you know, to see before with your product either that everything works and you can fully committed, that you can go with the next step.

00;18;09;04 - 00;18;35;18
Unknown
So this integration process is complicated, I would say, and we are trying to simplify it as much as possible. But you need couple of iterations. Let's say you have your product field, you see what doesn't work, then you iterate. We are calling it like data simulation one, data simulation, data simulation to three until everything is as expected and you can go with the next step.

00;18;35;20 - 00;18;53;12
Unknown
Little by little. And to go to like to do live environment now before a coffee break comes to an end today. Is there anything you want to share with our listeners that we haven't covered yet? I was preparing a bit to answer a question about the future because I thought that you would ask, So how the future looks like.

00;18;53;14 - 00;19;20;07
Unknown
So how does it look like? I see that there is like a lot of FOMO about all of the lands and, you know, a lot of software connected to our alarms. And usually it's very expensive to show this little shopping assistant because it consumes a lot of tokens and and there are no real examples at this moment. So it's like a lot of maybe a couple of them, but it's still on very, very early stage.

00;19;20;09 - 00;19;50;27
Unknown
And I see that merchants, these have this FOMO that they would miss something. So my feeling is that it would take much longer. And we we expect, I think, that, you know, we need to see what was happening with the Internet abduction. It took, you know, 50, 20 years to have real examples, which works, you know, So it means that we are all Internet users and we are getting value out of it.

00;19;51;00 - 00;20;19;26
Unknown
So now, yeah, we can go to Gemini and and receive ads, but still with all the software which is integrated with all those alarms, we still need to wait because it's extremely expensive. It's still slow, it's getting better. But when everyone when was talking, you know, 20, 22 or 2023, that, you know, in one year everything is going to changed.

00;20;19;26 - 00;20;48;01
Unknown
And now we are on 2026 and we still see that we need some years, you know, to to break that point. So I would say that we need to have this FOMO. We just need to know to wait until our systems would be prepared for that. So that's my assumption. Let's see now, have you what would happen in a couple of years?

00;20;48;03 - 00;21;16;22
Unknown
I value that because you're close to the source. You know what's happening in the AI and coming from that perspective, it's probably a good prediction to slow it down, wait and see what's happening, and just be informed all the way. Where can people go and find out more about Looper Search you Yeah website looper search that com we are trying you know to educate people to write blogs what is happening with search so usually everything is on our website global search dot com.

00;21;16;24 - 00;21;33;19
Unknown
Okay I will put the link in a show notes then you are just one click away. But thanks so much for giving us an overview of how important search is for every merchant out there and how important it is to get all the data together in one point to really get the most out of it. I hope a lot of people will reach out to you and I hope we will talk soon.

00;21;33;20 - 00;21;37;06
Unknown
Thanks so much for your time today. Thank you, Alex. Thank you for having me.