Boost Your Shopify Sales with Google-Powered AI Search — Sebastian Hooker | Innovations in Retail Technology, The Impact of On-Site Search on Ecommerce, Google Merchant Data Integration, Retail Revolution with AI, Predictive Analytics in Ecommerce (#277)

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 @clauslauter 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.

Claus Lauter [00:00:00]:
This is episode two seven seven of the ecommerce Coffee Break podcast. Today we find out why adding a Google powered search to your Shopify store is a good idea. Joining me on the show is Sebastian Hooker, co founder and managing partner at nimstrata.com. So let's dive right into it. But before we get started, a big thank you to our sponsors for supporting today's episode.

Claus Lauter [00:00:22]:
Loyalty Lion is a leading loyalty platform help Shopify brands rocket returning customers with their own loyalty programs. With Loyalty Lion, it's quick and easy to create a program that uses points and rewards to engage shoppers, secure more second purchases, and drive up lifetime value. Loyalty Lion has been increasing returning customer rates for thousands of Shopify stores for over ten years. Visit loyaltylion.com to find out more. Find the link in the show notes.

Claus Lauter [00:00:47]:
Have you heard about Partner Hero? They're experts in support on the ecommerce industry, known for their outstanding team building skills. Their main pillars, quality people and culture makes them a great fit for your company. Learn more on partnerhero.com or click the.

Claus Lauter [00:01:00]:
Link in the show notes.

Claus Lauter [00:01:03]:
This is.

Voice over [00:01:04]:
The e commerce Coffee Break, a top rated Shopify growth podcast dedicated to Shopify merchants and business owners looking to grow their online stores. Learn how to survive in the fast changing e commerce world with your host Klaus Lauter, and get marketing advice you can't find on Google. Welcome to the show.

Claus Lauter [00:01:29]:
Hello and welcome to another episode of the e commerce Coffee Break podcast. Today we want to talk about Google Search AI and how you can combine different things there to make a better Shopify store to get the most out of it. Very complex topic and we dive into this with the co founder of nimstra.com, Sebastian Hooker. He is the co founder and prior launching to Nimstra, Sebastian worked at Google as a cloud engineer. He began a new entrepreneurial journey when he spotted an opportunity to democratize AI, bringing Google Cloud Discovery AI to e commerce platforms at scale, starting with Shopify. So let's dive right into it. Hi Sebastian, how are you today?

Sebastian Hooker [00:02:06]:
Hey Klaus. Thank you. Doing well. Thanks for having me, Sebastian.

Claus Lauter [00:02:10]:
AI, Google Cloud Discovery AI, that's a mouthful and lots of our listeners will not know what it is. That includes me. So give me a bit of an overview, what we're talking about and how that will help Shopify merchants to boost their business.

Sebastian Hooker [00:02:23]:
Absolutely. So when most people think of Google, they think of search. Most people still go to google.com to find what they're looking for. There's an entire industry around SEO, around ads, marketing. Google is still very, very good at search. And what Google Cloud Discovery AI is, is it's a product that brings the power of Google search to e commerce websites. So Google does have a cloud division called Google Cloud. They compete with AWS and Azure.

Sebastian Hooker [00:02:51]:
And I think they really had to say, how do we appeal to retailers? How do we stop retailers from spending money on AWS or going to Azure? So they had to create these differentiated offerings. So what they did is they created a product and it's been renamed a few times, but underlying the whole thing is what's called the retail API, and it powers cloud, retail search and recommendations AI. And it is an API that allows you to import your product catalog to Google Cloud along with what's called user event data. So think of page views, adding to cart checkout conversions, things like that. And it learns from all of that data and provides AI powered search and recommendations on e commerce storefronts. So some major retailers you've likely heard of, like Macy's and Lowe's, have both used this product. It is implemented on their websites today. If you do go and search on one of these websites, that is Google's search and aipowered product discovery, powering that experience.

Sebastian Hooker [00:03:48]:
So that's kind of an overview of the product.

Claus Lauter [00:03:51]:
So you get the 300 pound Gorilla Google search into your small medium enterprise Shopify store. Now tell me, what's the downside if my search is not completely optimized in my store, what kind of impact does that have on my business?

Sebastian Hooker [00:04:05]:
We really look at conversions from search when we're trying to see if a customer would be a good candidate for this product. So a lot of times when users are searching on a website, they don't really know exactly what they're looking for. And product catalogs don't always know what to show. So if you are a retailer who sells clothes, you might sell women's clothing, and what customers might do is come to you and type in red dress in the search bar and they're looking for a red dress. What traditional search tools do is they look at your product descriptions, your titles, your meta fields, and kind of count. Like how many times does the word red show up? How many times does the word dress show up? And we're going to calculate that up, maybe do a little bit of matching and put that in the top left result. But what Google's product does is slightly smarter than that. I say slightly, but it's dramatically better than most off the shelf products and even better than some of the best in class ones out there.

Sebastian Hooker [00:04:56]:
You've probably heard of. But what they do is they say, okay, when people search for a red dress, what do they actually purchase? Do they purchase red shorts? Do they purchase yellow t shirts? And what they can do is take the intent of some of this behavior and return results that are most likely to convert. So sometimes in early implementations, we see results that don't look like they make a lot of sense, but they actually are very unique to the way customers use that specific product catalog and how they use the search. And if you have a poor search experience, you're going to have fewer conversions from search, fewer revenue. I think the obvious pieces are there, but what happens if you're a larger brand, or maybe even a luxury brand, is consumers view you negatively. So if ecommerce is a huge part of how you drive your business and you offer a prime in person retail experience, but frankly, your website experience leaves a lot to be desired, people actually remove their opinion of your brand from their hierarchy there.

Claus Lauter [00:05:51]:
Makes perfect sense. Obviously, as you said, stores with a huge catalog, and that's also what comes to my mind is like if you have a typo in there, the search engine will not find what you're looking for because they were looking for the product description and exact match. Now, how much is AI involved in your solution?

Sebastian Hooker [00:06:08]:
AI is a huge part of the recommendations AI solution, a smaller part of the search solution, but still very powerful. I tend to be careful with those two letters, AI. It feels like every company is trying to rebrand themselves as an AI company these days. And the reality is when we set off to do this, we don't actually compete with Google. We help merchants on Shopify and other e commerce platforms adopt Google's tooling. Because frankly, Google is better at AI than we are. What we are really good at is building very, very fast and effective middleware so that users can get value out of these AI platforms quickly. So AI really does power all of the Google side of things.

Sebastian Hooker [00:06:47]:
We focus less on the AI side and more on helping activate AI. But one of the products as part of this retail API is called recommendations AI. And at the core of it are models that are trained on your browsing data to return items such as frequently bought together. Others you may like similar items. And all of these models use data that's trained on your user events to facilitate the items that are most likely to convert. And you can actually choose. If you want to focus on conversion rate, click through rate, maximizing average order value. You really get to pick the output of the model that you're looking for.

Claus Lauter [00:07:22]:
With these tools now in the past, or if we're working, for instance, with Google Merchant center, you had to have a data stream which is formatted in a specific way so that Google could read the data, whatever. How does that work with your solution? Is there any kind of homework that I need to do from my Shopify site before it works, or how does it work?

Sebastian Hooker [00:07:39]:
That is an excellent question, Klaus. And there are a lot of feeds out there that move data into Google Merchant center, and I'm sure you've seen that Google is very opinionated on how they want their data. I would say the retail API goes a step further in the opinionated nature, and that's the intimidating part. And that's why for these major retailers, you have to pay a lot of money to have this implemented because they go upstream to your ERP, to your PIM, to different systems that you have in place to extract this data, transform it, move it into Google Cloud, and then keep it active in real time. So if you sell out of a product, is that making it over to Google Cloud quickly? And that's really the value proposition of what we've built, is that we've done all of that for you. So we have a single kind of set it and forget it mentality where you define your catalog schema. Once you tell us, for example, okay, Google's retail API wants to know what the brand field is and Shopify, is that your vendor? Do you use a custom meta field for that? Just tell us once. And then on all of our imports, we map that data for you and then we keep it in sync in real time.

Sebastian Hooker [00:08:42]:
And the only time you have to go back and edit that is if you add custom metadata to your catalog or you add product colors to your catalog.

Claus Lauter [00:08:51]:
Earlier you said that the algorithm needs to learn a little bit about your business, on what you offer. How does that work? How long does that take?

Sebastian Hooker [00:08:58]:
Yeah, so this really depends on how much volume of traffic is coming into your store. So to be fair, this is not a product for every store. If you don't have, give or take at least 1000 skus and at least 10,000 visitors a month, these models are not going to have enough training data to really add value. But the more you use it, the better it gets. So depending on when we start training the models and when we start actually using them, this can be as soon as a week, it can be as long as a month, depending on the site and the amount of traffic coming through. But generally speaking, the search results are going to be miles ahead of an existing solution on day one and they only get better from there. So the idea of okay, yes, intent based search works very well on this product catalog. And now as we combine it with the actual purchase behavior, we see improvements.

Sebastian Hooker [00:09:45]:
But one thing we do to combat this if you've been collecting this data already, if you have Google Analytics in place and you have this user event data, what we can do, it's a bit more of a manual process today depending on where that merchant keeps this data. But we can upload historical data as training data so you don't have to wait from the day of installation to train the models. You can actually use your former data to onboard here and get value out of that much more quickly.

Claus Lauter [00:10:11]:
And now a quick break to thank the sponsors of today's episode. Tired of watching customer drift away after their first purchase? Then it's time to rocket your returning customer rates and drive six times more second purchases. With loyalty line as a leading loyalty platform in the Shopify space, loyalty line exists to make it as easy as possible for you to launch and manage your own loyalty program. They'll even build your loyalty page for you. With loyalty line, you unlock everything you need to engage shoppers with points and rewards and integrate loyalty data into your email, SMS and help desk platforms.

Claus Lauter [00:10:42]:
You also get a five star support.

Claus Lauter [00:10:44]:
As shown by the customer reviews on the Shopify App Store and G two. Visit loyaltyline.com to find out more and start your loyalty journey. Find the link in the show notes.

Claus Lauter [00:10:54]:
As a prominent player in the e commerce support arena, partner Hero specialize in delivering personalized customer experience solutions with a clear focus on helping you not just meet but exceed your goals and requirements. They have become experts in e commerce support. Their commitment to tailoring solutions to your unique needs ensures that your support experience is nothing short of exceptional. Partner Hero is more than a CX company. They are your partner in success. Visit their website and learn more on partnerhero.com.

Claus Lauter [00:11:20]:
Okay, that's a huge advantage. I mean, often it's like you start from the moment when it gets activated and then you need your four weeks, six weeks, eight weeks. But if you can start with historical data, a huge benefit there. You mentioned event tracking. You mean check out add to cart? How does that come into play?

Sebastian Hooker [00:11:36]:
That is again all built into our app. That is where the retail API does have opinionated decisions on how you pass that data, in, what product ids you send, what type of page you're on. So we use the construct of the app embed blocks in Shopify it truly is two clicks to start tracking events. Once you have our app set up and you go into our theme installation setting, you just click enable the app embed and you enable the app pixel. And what these two things do is they work together to identify which market the Shopify customer is coming from. Because if you have a canadian store and a us store, those users might have very different behavior. They might buy thicker socks up in Canada in the wintertime, and the frequently bought together recommendations need to reflect that. So what our user event tracking pixels do is it first identifies what Shopify market you're coming from, and then it sends the data to the correct backend place in the retail API so that the training data is using accurate location data as well.

Claus Lauter [00:12:34]:
Let's talk about implementation. Is it you're in the Shopify App Store, what are the steps to get started? What's the onboarding process?

Sebastian Hooker [00:12:41]:
We are now a public app in the Shopify App Store. When we were deciding whether to build this, we kind of thought, okay, should we do a private app and one time implementations? That's kind of what the industry does today, given the complexity here. Or is it possible for us to make this so easy? We can come out with a public app and kind of do minimal live chat or email support for onboarding and then offer maybe up to a premium onboarding experience if you want to look at some of these historical user events to bring in or map. So today, what the onboarding experience looks like, it's completely self guided. You go to the Shopify App Store. Our app on Shopify is called Nimstrata Connect. They didn't want us to use retail cloud Connect. They said it was too generic.

Sebastian Hooker [00:13:21]:
So Nimstrada Connect is the name of our app in the Shopify App Store, and you install the app. And from there we guide you through setting up your Google Cloud environment. So when I worked at Google, we had a training platform called Quicklabs, and those labs were set up to be copy and paste this command, press enter, you pass the quiz, check out what you do. It was very easy to learn. So we took that same model and we said, how would someone learn to use this platform, and how do we build that user experience into our app? So you go through about six steps where you just copy and paste a few commands, and then we verify the connection between the Shopify store and Google Cloud. As soon as that's done, the hardest part is over. From there, you configure your catalog schema, you install two or three app blocks wherever you want to put the recommendations. This also works on collection pages, so it's not just search and recommendations.

Sebastian Hooker [00:14:09]:
You can use it for a collections page and completely eliminate the amount of merchandising a human has to do. So if someone clicks on a category or collection that data is fed by the retail API and again learns from historical behavior on how to position those items. So once that's done, it's really a matter of deciding which recommendations AI models you might want to build in the future. You might decide to a b test some things, but the reality is you should see an increase in conversion from search on day one and continue to see that grow. And as soon as you have enough data in the Google Cloud platform to create what's called a recommendations AI model, you can create that and dump that into an app block by just pasting what's called a serving configuration id. So we surface all of this within the app. It is a very guided experience, but generally speaking, this can be done from a cell phone if needed. Don't ask me how I know, but it's possible.

Claus Lauter [00:15:01]:
I think most of the traffic nowadays is coming from mobile devices, and that makes perfect sense to start there. Now, who's your perfect customer? Are there specific industries, niche, verticals where it would work best?

Sebastian Hooker [00:15:12]:
Yeah. So our perfect customer would have at least 1000 skus. If they really want to train their data quickly, we would want about 50,000 visitors per month. I do want to emphasize that this is an enterprise product, and I know you have a lot of listeners with smaller stores, maybe dropshipped, and we would love for them to grow into this product. But you really do need the volume of data for the models to be effective. If not, you're kind of paying for a product, you're paying for a Ferrari when you need like a Corolla. So it's not really worth it there. But when it comes to types, I think this is one of the most exciting pieces because this is such a new product from Google Cloud, they are still learning what types of catalogs perform best.

Sebastian Hooker [00:15:51]:
So we've seen a lot of success while I was there in the apparel industry looking at clothing, shoes, whatnot. But this was actually super effective at hardware stores. So when you think of like an electrician or a plumber, it's amazing what they can come up with to search. So a two x four x twelve in the search bar really confuses a lot of platforms. But what that is is a twelve foot long two x four, that's a piece of lumber. And what the search algorithms can do is say, okay, every time someone searched a two x four x twelve, they purchased a twelve foot two x four. Let's start showing that in the search results. So even if the product catalog isn't a perfect fit, the idea that it's constantly learning based on your site behavior could make it a good fit over time.

Sebastian Hooker [00:16:31]:
I think if you're questioning it, it's more worth like an A B test to say, is this better than whatever solution we've kind of tuned over the years on day one? And can we improve it even more after that just by letting people use it?

Claus Lauter [00:16:45]:
I can see that specifically you said hardware store. One example that comes to my mind. I spoke to a merchant a while ago. He's selling parts for Westbus scooters, for Westbus scooters of the last 50 years. So he has like a million different skills, and I think that would be a perfect fit for him. Tell me a little bit about the pricing structure. How do you work with that?

Sebastian Hooker [00:17:03]:
Yeah, so today we do have a floor of $99 a month as a base fee. And this is really intentional to make sure that people really want a solution like this before adopting it. When we do look at that enterprise customer base, kind of the corporate sector that goes like one step above a small business who has just started out, we would want them to pay that $99 fee to make sure that they're opting into this product. And what that fee covers is, frankly, all of the middleware that we had to create to return these results in real time. So Shopify actually runs on Google Cloud. The retail API is a Google Cloud product. And then we built everything to interact with both of those platforms, essentially seamlessly, so you can deliver search results in fewer than five milliseconds on a site. So our goal was to say, how do we make this a faster, more powerful solution than if someone went and built this themselves? And with that as kind of our North Star, it led to a very cool microservices based architecture, where we return these results very quickly.

Sebastian Hooker [00:18:04]:
But to support that, we do need to charge a fee for that middleware. And what that does is it keeps catalogs up to date. So we're running imports constantly to make sure catalogs are there. And then after that, we just charge like Google. So Google charges $2.50 per thousand API calls. It's not much. We charge $0.50 on top of that for a total of $3 for recommendations, it starts at, then we charge a nickel on top of that. So we do have similar volume based pricing that Google has.

Sebastian Hooker [00:18:34]:
And then what we can also do is work with customers to predict their bills. So, for example, if they only want to use something like this for search, it's going to be different than if they also put this on all their collection pages and deliver recommendations. Eventually, you could find yourself in a point where you're making an API call to Google on every single page view, maybe multiple. But if you can see the business case of, okay, we convert 5% more using these tools, we'll work with customers to make sure that the pricing aligns. And just being a new business, we have to figure out, okay, what SKU gates do we want customers to see on this pricing? So right now, our enterprise plan starts at 15,001 skus. Now, that's not to say if a customer came to us and said, I have 16,000 skus, I don't need your enterprise support. Can you work with me on a cheaper plan as this one on one kind of onboarding experience? That is part of our conversations. But generally speaking, we price it in line with Google Cloud.

Claus Lauter [00:19:32]:
I can see that. And specifically, I think most user journeys nowadays start with a search. There might be occasionally people were just browsing, but I think a lot of people are using search. And the better search is, the higher the likelihood is that you get the conversion. Before we come to the end of the coffee break today, is there anything that you want to share with our listeners that we haven't covered yet?

Sebastian Hooker [00:19:50]:
I think at a really high level, knowing where your conversions are coming from and where that can be optimized is important for any merchant. We were very surprised at some of the increases in performance on stores that we didn't expect. So if you want to run an A B test with us, that's something we can do. Right now, we're covering the cost for 14 days. And that's not just our cost. We're actually covering Google Cloud costs as well. So if a merchant is interested in kind of seeing if this would be effective on their catalog, that's something that they can come to us and do. We are only accepting a couple more for our December cohort.

Sebastian Hooker [00:20:24]:
However, this will be picking back up in January. But, yeah, that is what I would leave listeners with.

Claus Lauter [00:20:29]:
Okay, excellent. Sebastian. Where can people find out more about you guys?

Sebastian Hooker [00:20:32]:
Yeah, you can find us@nimstrada.com. That is the easiest place to find our website if you google nimstrada. I'm glad to say we've come up first now and then on the Shopify App Store. Our app is called Nimstrada Connect. You can sign up there on every single page of the app. There's a get help button in the top right. You can click that. You can live chat with us.

Sebastian Hooker [00:20:50]:
You can set up a calendar video meeting if you want to do a guided onboarding over screen sharing. We really do aim to kind of iron out any kinks in an app. We are a new app, so any feedback from there we take very seriously. And for any merchant looking to add this to their website, we really want to offer a premium experience.

Claus Lauter [00:21:09]:
Okay, sounds good. I will put the links in the show notes then you're just one click away. Sebastian, thanks so much for your time today. I think it was a very good overview of how powerful search is and that you really should squeeze out the most out of it, and I think your solution helps with that. Thanks so much for your time today.

Sebastian Hooker [00:21:24]:
Thank you, Klaus, thanks for having me.

Claus Lauter [00:21:26]:
Hey Klaus here, thanks for joining me on another episode of the e commerce Coffee Break podcast. Before you go, I'd like to ask two things from you. First, please help me with the algorithm so I can bring more impactful guests on the show. It will make it also easier for.

Claus Lauter [00:21:38]:
Others to discover the podcast, simply like.

Claus Lauter [00:21:40]:
Comment and subscribe in the app you're using to listen to the podcast and even better if you could leave a rating. Secondly, please take a moment to check out today's episode sponsors. They play a crucial role in keeping the show and our newsletter available to you for free.

Claus Lauter [00:21:53]:
Thanks again and I'll catch you in the next episode. Have a good one. Before you leave, don't forget to visit.

Sebastian Hooker [00:21:58]:
The sponsor of today's episode.

Claus Lauter [00:21:59]:
Have you heard about partner Hero? They're experts in support on the e commerce industry, known for their outstanding team building skills. Their main pillars, quality people and culture makes them a great fit for your company. Learn more on partnerhero.com or click the.

Claus Lauter [00:22:12]:
Link in the show notes Loyalty lion is a leading loyalty platform helping Shopify brands rocket returning customers with their own loyalty programs. With Loyalty lion, it's quick and easy to create a program that uses points and rewards to engage shoppers, secure more second purchases, and drive up lifetime value. Loyaltyline has been increasing returning customer rates for thousands of Shopify stores for over ten years. Visit loyaltyline.com to find out more. Find the link in the show notes.