Latest / Talk Commerce / PIM Systems Are Revolutionizing Agentic Commerce with Pimberly's Martin Balaam
Transcript
- Brent Peterson (00:01.373) Welcome to this episode of Talk Commerce. Today I have Martin Balaam He is the CEO of Pimberly. Martin. Go ahead, do an introduction for yourself. Tell us your day-to-day role and maybe one of your passions in life. Martin Balaam (00:13.004) Yeah, hi Brent. Thank you for having me on. My name's Martin Balaam I'm the CEO and founder of Pimberly. We are a SaaS, PIM, product information management and digital asset management platform. We founded, I founded the company in the UK, but I've been over living in the US for the last three years as we expand into the US. So that's who I am in terms of my passions. As well as my family etc. I love classic cars and I love to Tinker around with those and my wife says I've got problem. don't know if you'll get this joke, but I've probably got the most expensive scrapyard There is as it were so that's me Brent Peterson (01:00.115) I can say that 20 years ago I probably had three sitaruns and at least four or five minis in my yard. So I totally get it and I imported a whole container full of the classic mini into the US about... Martin Balaam (01:06.588) wow. Martin Balaam (01:12.598) Yeah, well that's obviously, given the experience I am, that's the one that I would be thinking of when you said Mini. Yeah. Yeah. Brent Peterson (01:20.071) There you go. Not the new and the old one. Yeah, anyways. So, thank you for that. Before we get started, you have graciously volunteered to be part of the Free Joke Project, and now I should have had a car joke. anyways, here we go. I said to my wife, you seem a lot happier. She said, thanks. I uploaded our entire argument to ChatGPT, and it said I was absolutely right. Martin Balaam (01:47.389) Yep, I'd give that a 10, 11, that or a 13 Brent Peterson (01:51.441) All right, that's perfect. All right, Martin, so I think most of our listeners know what a PIM is, but just give us the elevator pitch when somebody walks in and says, what's a PIM? What do you tell them? Martin Balaam (02:04.556) Yes, so a PIM essentially is an e-commerce tool platform that is your single source of truth for all things product data. And data is, know, whether that's just data or whether it's images, videos, documentation that might need to go with it. And what a PIM does is it basically accelerates the time and optimizes that product data for all the end marketplaces and websites and whatever that you need to sell at, regardless of... what country they're in, what language they need to be in, et cetera. And so it really helps you on board product a lot quicker, get more optimized, more engaging product information, and ultimately allow you to sell more and to convert more. Brent Peterson (02:49.683) I know you kind of hinted at it that all the different endpoints that we have now to sell products. think a lot of people, maybe 10 or even 15 years ago would have thought, you know, my, my e-commerce is a great PIM, right? But now maybe kind of explain to us why that central point of having all your product data and then being able to publish it to all the different endpoints is so important. Martin Balaam (03:18.86) Yeah, absolutely. On that point that you said, I'm sure people think e-commerce, as you said, 10, 15 years ago, they think about a website. Well, nowadays, when you're wanting to sell, because of true globalization, can literally, even if you're selling physical products, you can get those products pretty much next day anywhere in the world. So your opportunity is being able to reach out to as many consumers in the globe as you like. And obviously those consumers are just like you are, they have the go-to e-commerce platform, whether that's an Amazon, an eBay, a Tmoo or whatever. Or if you're in Holland, it might be a Ball, et cetera. And if you want to sell your products around the world globally, you probably need to be touching... at least a hundred different e-commerce platforms, marketplaces, call them what you want. And really, if you think about as you go and expand outside your normal geo, you're increasingly going to need to get your products listed onto all these places. And every single one of them have a different methodology of how to get your product data on there, how to interact with their systems. And so what's become essential for people there and businesses that are actually doing global e-commerce is to actually have a simple source of truth, single source of truth for product data, which is a PIM that can actually manage all these different variants, understand what the different validation rules are for each of the endpoints and be able to deliver the validated product data to where it needs to be such that it gets listed and you get to sell. Brent Peterson (05:13.171) Are there any particular struggles or anything that people should now be looking at when they're PIM? When they're thinking about what is perplexity going to want or what is CHP going to want when we're starting to do the LLM purchasing? Martin Balaam (05:30.903) Yeah, I think, again, for those of those that have started to plan or even implement for, you know, this sort of agentic LLM new landscape that we're on, actually, you know, creating a product data that's that's listed into your e-commerce website or etc. is just one element. That's the human side of it. And nowadays, if you really want to be sort of pushing the envelope to the LM, you're also actually publishing that product data in a machine, easily machine readable format. And people say, well, what does that mean? Well, it's, for example, it's making sure that all your data is in accordance with things like schema.org, that you're actually publishing an LM sort of friendly product data feed, ideally with... intermingled with what stock you've got available where, at what time. Because when LLM agents or whatever are interacting with that, it's usually very conversational and they're wanting to give a response back to the consumer that's specific to where they are now, what stores are they near. Because they're on their mobile phone, they know exactly where they are, they know whether they're walking or whether they're traveling in a car or a train or whatever and where they're going to. So they want to give a really relevant response back, which means that you have to give better raw data that's in that structured content that that LLM can understand and understand precisely. Brent Peterson (07:19.731) I hosted a panel at ETale, is a online retail conference that happens in the spring and the fall, whatever. But one of the things that we asked the audience was how many of you as merchants are employing agentic? Like that's the buzzword, Agentic commerce. Nobody raised their hand. So it kind of tells me that people are planning. Martin Balaam (07:39.436) Yeah. Brent Peterson (07:46.557) but not implementing yet in terms of agent at commerce. And I think that there's all wide variety. There's a tool set that you get from all your product data that you can actually create a bot, right? That is a agent. But then there's this new idea of deploying a bot to go and buy something for you. How far down the road do you think we are from that happening? Martin Balaam (08:04.813) Mm-hmm. Martin Balaam (08:09.326) Well, I was recently at a distribute, I think in B2B world, a lot nearer than what people think. In fact, actually, I think in B2B, they're probably, you know, unusually might be overtaking B2C in terms of adoption of some of this technology. Because obviously B2B, the primary objective is buying the precise product that you need. the secondary objective is sort of like, you know, is it the right price? Because in B2B land, can always, if you feel that you've not been given the right price by your supplier, you can always go back afterwards and negotiate something on that. But fundamentally, if you're needing a part for a piece of machinery that's broken, the number one thing is, I need this, I need that, I need the exact specifications. And if you don't have that part, actually in stock or maybe it's end of life, what is the next nearest replacement that will work? And that is a lot of very specific data that lends itself to a Gentic. I would suggest that certainly from the conversations I've had that that's already getting deployed in some of the more mature B2B companies. Whereas in B2C land, think, I don't know, I think the general sort of thing is that consumers are using whichever LLM they like best, whether that's chat GPT. or whether it's Gemini or whatever, and they're asking it questions and asking it to do things for them in terms of maybe write a complaints letter or write this, that and the other. But they're not yet asking it to or giving it enough trust to go away and transact on their behalf. But that is probably not too far away, especially for the very early adopters. And especially if the likes of OpenAI and whatever, you know, start getting some sort of banking arrangement in place. And I think you might see consumers starting off with like micro payments sort of stuff, saying, well, actually, I don't care. You know what toothpaste I want, or you know what toothbrush I use or whatever, and it's only a $5 thing. I'm happy for you to go in and buy it and go and you know where I normally, I like shopping at Walmart or I like shopping at Amazon. Just go off and order me that. And I think that will start coming pretty soon. Brent Peterson (10:28.831) Yeah, I think you put you I mean, I'm going back to B2B. I think that's more more of an interesting play. And I think that especially for PIMS, allowing an agent to come in and get all the product data that it wants. And I think if we look at what an ERP does or WMS ERP system is usually the central point for financial data and maybe your core product data. A lot of I think a lot of merchants who are maybe bootstrapped tend to try to put as much into the ERP or the e-commerce engine, but your real, real beautiful data, we'll call it, happens in your PIM, right? And do you see that the PIM might be the first point in which an agent would come to you and say, hey, I'd like, I need to know everything about this widget. Tell me everything about it so I can go back and then, then we'll assess pricing later even. Martin Balaam (11:15.95) Yeah, no, don't forget, these agents also understand, just like Google, at the end of the day, it will also be scoring. It'll be saying, how often do I come to this supplier's PIM and get exactly what I need? And if it's actually, it's pretty good data. And don't forget, this data might be locked in a PDF document, especially if you're looking at industries where the products are evergreen. and literally you could have a, again, a heart back to components and things like that. You might have a machine that's 20 years old and actually the product isn't existing in a data form, but it is existing buried in a PDF service manual or whatever. Then again, suppose you're that LLM know, well, okay, and then obviously you go away with your little bots in your PIM to say, I need to go away and find this, have I got it? And giving it back. Yeah, I I absolutely see it going that way. And also, as I came back at the beginning of that question, it's if you can deliver that, guess what, the box would keep coming back to you because all they want to do is do their job. They want to get told they've done a great job and then move on to the next one. Brent Peterson (12:33.533) Yeah, and actually being from the Midwest and using a lot of Claude code, I can tell you that the LLMs don't care if you're passive aggressive or not. just want to be direct and you can swear to this as much as you want and they don't really care. Martin Balaam (12:42.446) Yeah. Martin Balaam (12:47.756) Yeah, as well, know, I think there is a big difference, you know, as I said, in B2B and B2C, B2B would be more about precise long tail data. I think B2C would be more contextual about, you know, and that's a big thing for, for brands in B2C is to start collecting as much contextual information about the products that you sell. because it's more about the story of the product. I think that from the LLM's point of view, the consumer will be asking it. They'll be talking more about what they want and where they're going and why they need it and how they want to feel. And so you don't get that from raw data. You get that from the stories behind the other people that have bought that product and making sure that you're starting to collect that and this is where reviews are going to come back in and I've already seen a lot of B2C customers talking to us about, I need to get my reviews in my PIM now. Before it would be it's in my review engine, we don't put it anywhere near. Now it's like, no, no, these are really super important now. Brent Peterson (14:03.379) How, you mentioned PDFs and I actually, talked to a company that was looking for investors for a PDF to data extraction for specifically for procurement and B2B. And a lot of older legacy systems are either giving or sending a PDF instead of EDI, which is even older probably than PDF, but. How important is it that you see even the vectorization and then that data is in a database instead of in a PDF? How important is that in the B2B space? And how much is Pimperley looking at the vectorization of all these flat files? Martin Balaam (14:50.902) Yeah, I mean, we think it's really important. And in fact, we've got a research project going with a university in England to do just that. because at the moment you can do it at low volumes, you can use AI to pull out the data that you need. But you still need a lot of human interaction to set up all the prompts so that it is super accurate. And really what you want to be doing is, what we're looking at is, we've got, let's say 20, we've got some customers with 20, 30, 40 million pages of locked up data at the end of the day. And what you want to be doing is a multi-layered sort of AI approach, which is sort of mimicking what you would do with humans if it's doing the prompts and doing the QA and whatever. And we've got, so we've got a project going on this year just specifically to do that. So. Yeah, we see that as super important for our sort B2B customers, especially in industrials, et cetera, which as I say, very different to the B2C where it's capturing all that conversational interaction with their product that they're after. Brent Peterson (16:09.779) In the green room we talked about a couple of different PIM systems. What sets Pimberly apart from some of the other PIMs out there? Martin Balaam (16:19.47) Yeah, I mean, we're one of the things I guess that's really different about us is where we come from, which is fundamentally I have a family business that manufactures and distributes furniture in the UK. so Pimberly was being built around that sort of ethos of, you know, complex multi many sort of combinations of products, etc. being able to deal with some fairly hefty digital assets that need to go with it and a lot of conditionality and complexity. I think if somebody's, what I normally do is really sort of say, if you've got great product data but your issue really is how do I syndicate that onto multiple sort of retailers, et cetera, et cetera, you know, some great syndication tools for doing that. So I'd probably put. put hand up and say that's probably not Pimly. But if you're looking for a Pim down solution that can really truly be the single source of truth for all your product data, handle lots of high complexity and lots of conditionality, and also you care more about what the product data is and how refined it is and specific to your end use point, that's us. And then the other thing is that we're sort of, I call it as a full service software company. know, everything is done in-house, including the implementation, you know, so, you know, rather than handing off to a third party integrator, which we do work with, you know, obviously if it's part of a big digital transformation, you do, but fundamentally, you know, it'll be our team, our team, team that do nothing other than implement PIMs and probably, you know, I'd say probably more Pimp Solutions in the world than any other dedicated team. So we know what they're doing, they know what the pitfalls are. And also, because we're British, we'll fundamentally, we'll say if we can't do something, we won't oversell. Brent Peterson (18:34.887) Right. I have to ask the question about tariffs and how much that's impacting the amount of data somebody has to keep. And I think as you get more granular and as you have to find different suppliers, maybe some of that data has to change. Do you see an advantage of having all your data in that one spot that even during some of these times where you're doing a lot of shifting, especially if you're manufacturing something and you're trying to keep those products in one place that you can shift it and then I guess the finally on that little bit what do you also store the manufacturing data of the whole part as most of that happened within the ERP or the WMS system? Martin Balaam (19:18.446) Yeah, so on the first point, obviously, if you suddenly have faced, I mean, we've had tariffs for a long time, but we've not had them change as much as this, as they have done for a very long time and be as impactful as well in terms of cost, potential cost. So one of the first things obviously that you do when faced with potential significant increased input costs is ask, well, how does it impact me? which of my suppliers are going to be impacted. don't forget, you've got to look down a couple of layers because, okay, there might be a US supplier today, well, right now, but if two thirds of their components are getting hit by tariffs, guess what? Even though you might have that purchasing power to ultimately say, I'm not taking a price increase, at the end of the day, you know that medium long-term, if that's a not sustainable partnership for them with you. that ultimately you'll have to ultimately find a new supplier. first of all, having that information and then, know, which has been really, we've seen a lot of people, you know, you can see, you know, adding additional attribution, et cetera, to capture this so they can understand. And then the next thing is being able to move your, you know, your supply chain around, you know, in terms of, well, who's got what, where's it coming from, what impact does it has? It's certainly been a... a benefit to our customers having that data and if they're not, we've seen them adding it. The other thing as well is that especially you get, know, suppliers with huge volumes of product and product items and variants and whatever, some of the things that they've been faced with, just the huge processing that's needed to compute, to actually run and say, okay, well, what happens if this happens? What happens if that happens? What happens if that? What if we move from here to there? You know, I think... I would imagine there's a lot of AWS compute and whatever that have been very happy with all these tariffs because of all the extra compute that's been going on to work out all the different scenarios. that's, and we've seen that with a few customers where it's like, well, you know, we, because they're having to cut it at their entire database sometimes. And, and they're saying, well, look, it's taking me, you know, nearly all night to get to this sort of thing. And then guess what? changes the next day and they have to run it all again. So, so that's, that, that's, that, that part of it. Martin Balaam (21:41.902) In terms of information in the ERP, etc. and WMS, we tend to find that where people have a PIM, they will consciously try and then minimize the amount of data that is required in the ERP so that really it just is for being able to transact. And also don't forget if... Again, when you come to personalization and even if it's B2B or B2C, if you're wanting to display a different product description, different features and bullet points for exactly the same skew, then you don't want to be putting that in your ERP. don't want to, I mean, you come across regularly people when you're implementing a PIM for them, they'll have multiple different skews for the same product. Just because they've had to do that to be able to get... different product descriptions for different customers, as it were, or different marketplaces. And that's definitely a thing that we try and help them with and say, no, you don't have to have all that. We can go back to having one SKU, not having to have somebody when they get an order through, having to say, well, actually, no, that is the same SKU as that one, so with a note on it or whatever to make sure they understand. So yeah, and then warehouse management systems similarly. you know, again, helping them reduce the number of, if you like, phantom skews that they're having to use because they've been needing to manipulate the system to give them what ultimately their marketing team and sales team wants at the customer facing them. Brent Peterson (23:25.137) Yeah, wow, you're bringing back great memories, Martin, of my previous days in that space. So we have a few minutes left, Martin. Before we go, I give everybody a chance to do a shameless plug about anything they'd like. What would you like to plug today? Martin Balaam (23:43.106) Yeah, well, I've got to sort of plug Pimberly out now, if that's where I grew on. So yeah, mean, if we've got listeners out there that fit that bill, you've got complex product data, you may be setting an old legacy on-premise MDM or Pim solution, and you're looking at what maybe the next step is for you, then you can go on our website, pimberley.com, lots of places on there to get in touch with us. I'd love to hear from you. Brent Peterson (24:13.509) And the show, do you do the show circuit? Will you be at any conferences or anything coming up? Martin Balaam (24:19.214) Yeah, yes, I'm at the We're at shop talk fall in Chicago next week a week after next sorry and Also, I'll be presenting at the National Association of all sailors Innovative tech innovators summit next week. So and then after that we're at the Hardy show, which is the HVAC showing in Las Vegas and so so yeah, we keep ourselves busy Brent Peterson (24:48.243) Excellent. Yeah, I'll see you at Shop Talk. I'll be there as well. So, all right, Martin, it's been such a great conversation. Yeah, thank you so for being here. Martin Balaam (24:50.742) Alright, look forward to it. Martin Balaam (25:01.038) That's a pleasure, thank you. I bet you wish you still had a mini left. Brent Peterson (25:04.478) That's right, I sure do. All right, that went great. Yeah, thank you.