Lift and shift is the wrong goal for a PIM move

Episode 15 - Lift and shift is the wrong goal for a PIM move - Transcript

Matt: Welcome to Data Versus Commerce, where we explore the messy middle between database and doorstep. I’m Matt Johnson.

Floyd: And I’m Floyd Blaikie. Let’s dig in. Okay, welcome back Willem, Jay, and of course Matt — it’s a familiar group here today, and I just have to say, before we even started this recording, I learned more than I think I’ve ever known about foam rollers. So thank you to our guests, one of whom is an ultramarathoner. I won’t say who — I don’t want to put anyone on blast if they’re not ready to talk about it.

Matt: It’s me.

Floyd: It’s Matt — and that is why he shaved his beard off, actually, it makes him more aerodynamic.

Willem: More faster.

Floyd: Yeah, but maybe that’s for a different time, when we start a new podcast about running. Rather than talk about running, we’re going to talk about PIM and making the move — this is our third episode of three. I want to say it now, and we’ll mention it again at the end: later today, in fact the day this airs, Wednesday, September 2nd, we’ll be joining Willem and Jay again for a LinkedIn Live event. So if you aren’t signed up and you’re listening on release day — thank you, first of all — look us up on LinkedIn, Data Versus Commerce, or look up Pivotree, or inriver, and we’ll all have links to that live event. We hope to see you there. We’ve spent two episodes on why people get stuck in the PIM conversation — what happens when they hit the ceiling, when legacy PIM or sales stops being good enough, the data foundation — and now we need to talk about the elephant in the room: how do you actually move? Who wants to get into it?

Willem: Yeah, I can get into it. A lot has changed — probably a year ago, everything was done without AI. Now AI is really front and center in delivery, and that’s making a huge change in the timing and cost of delivering a new platform for folks. We do so much with AI in delivery — needs gathering, creating user stories, grooming those user stories, going back and getting questions you’ve missed presented to you by the AI platform. The teams really work hand-in-hand with a tool plugged into the details of what a platform can do, and then you get build instructions out of those user stories, created in connection with an AI platform — it speeds things up tremendously. The build instructions create an initial implementation, you prototype it, get it in front of users. Post-delivery, it’s really helpful too — we create test scripts as we build, again using AI, and post-delivery support triage has sped up as well. So in all the facets of the software development life cycle, there’s now AI inserted, and you see speed increasing and cost decreasing. Tremendous change from how it was done less than a year ago.

Jay: As the old proverb goes, may you live in interesting times. I think, as part of writing the user stories, it’s also really interesting using AI to understand what the actual workflow looks like today — and it’s frequently not just what’s in the system. Somebody needs to understand that every Tuesday afternoon, Bob sends Mary an email with updates that need to go out through certain channels — that’s the kind of thing that can now be captured and built in as a user story. So it’s not just that what used to happen now happens somewhere else — it’s a fundamentally different workflow and understanding of the end-to-end process.

Willem: Yeah, and when you talk about replacing a platform, the interesting thing is you can take a configuration and read that into your AI platform, gather what’s happening today — but of course you don’t just want to lift and shift. There’s an element of wanting to radically do things a better way. You take a base implementation, look at it, use it, but then really improve on it, because things have also changed in the channels you now need to deliver to. Agentic shopping is a great example — you need to be ready for that, and create a solution that can support it.

Jay: We’ve seen across the spectrum, everybody, even B2B buyers, starting their research using AI, and even selecting products using AI. So the question isn’t do you need to get ready for it — you’re either going to be ready for it, or unfortunate things are going to happen, you’re going to be invisible. Willem, I like what you said — it’s not lift and shift, it’s lift and be prepared for what’s coming down the road in commerce and discoverability.

Matt: One of the things I was thinking about, as we’re thinking through the ways AI has fundamentally changed how companies stand up a PIM, implement it, integrate it — also all the adoption of the platform itself, and thinking through the business challenges around getting people trained, getting them to use it the right way, identifying where the gaps are, tracking the performance of what comes out of that adoption — that’s really interesting too. I was curious what you guys thought about the human resources involved in these PIM implementations.

Jay: I’ll take a bit of a “yes, and” approach to that, Matt, because part of any implementation is deciding what you’re going to have the humans do, and what you’re now willing to have agents start to take on, under the control and governance of humans. Where are you willing to have copywriters? Where are you willing to have translators aware of your glossaries and brand guidelines? Is that a step, as you look at a move to a modern platform, that can simply be replaced by what humans used to do, freeing people to do more high-value work, or review the implementations that have gone on? As I think many people would agree, content is free — any one of us could go into our favorite LLM and generate content today. Governed, underwriteable, on-brand, trusted, compliant, regulatory-approved content is not free, and that’s where you need to start building the infrastructure to get there. So I took a bit of a twist on your question of how you measure the value of the platform’s adoption by humans — it’s not only are the humans adopting it, but what are we now asking the humans to do in the new world order?

Matt: I love that.

Willem: I was going to lead more with how do we run projects, and what focus people come in with. What I’ve seen a lot in the past is the project team from a company deploying a solution would come in and listen. I think that’s changed a little — if you do this a lot, you’re not coming in without opinions. That doesn’t mean the customer’s opinion doesn’t matter, but in some ways, there’s a saying: you’re unique, just like everybody else — and that has an element of truth to it. So we can come in with defined, strong opinions about what should happen, and then, of course, there’s a lot of back and forth and nuance about what actually happens. But that strong opinion about what you should do is really important — you need a strong team, both on the architecture and project side, that says, “This is how you do it best,” and then you adjust based on what the customer needs. That, together with change management, makes for very successful projects — change management, communications, early review of what’s happening, getting the stakeholders involved, all of that is key too. So I don’t know if we answered your question, Matt, but we both took a stab at it.

Matt: No, you did a great job. What’s interesting is what Jay was saying — this is an opportunity. When you go to implement a technology like this, which historically has been an extremely manual, frustrating process where you continually want to throw bodies at it, but there’s really no bodies to throw — this is a great opportunity inside your org. You’re getting pressure from the executive team: “How are you implementing AI, where are you putting AI into our operations?” This is a great chance to say, “Wow, look at these manual processes — now that we’re implementing the technology, we can re-engineer the processes and figure out where our human resources can best perform,” and utilize their expertise and skill, getting them out of the spreadsheet world, spending their time as drones, and letting AI do the heavy lift while they’re strategic.

Jay: Watching AI take unstructured content — like the catalog your manufacturer still prints and sends you a PDF of — and turn that into structured data can be amazing, because the amount of profanity that has to come out of anybody’s mouth when they’re handed a catalog and told to type it in manually is probably not appropriate for the younger members of our audience.

Willem: Yeah.

Floyd: Like my seven-year-old, who knows I talked about her last time, and now wants to listen to every episode.

Matt: Yeah, and thank you for that, Jay, because Floyd is very proud of our clean rating on iTunes.

Floyd: That’s right, it’s true. Listen, I feel like if I’m the person getting the email every Tuesday, like you said, Jay, and I’ve got to go type in the data and make sure it’s formatted properly — I’m sold. I know the technology’s out there, I know it’s improved rapidly, I know it’s less expensive than it’s ever been, and I know it’ll save me time. But maybe I’m not the person signing the checks — in the organization, there are so many other people who maybe have been burned adopting something like this before. My sister-in-law, for example, is leading a CRM migration project right now, completely unrelated industry, she works in nonprofit healthcare, and I’m her designated therapist — I’m getting texts from her at 11 p.m. like, “Well, my last batch of 10 million records just failed.” What’s actually changed, in a way we can prove, between maybe your last terrible implementation of something that was supposed to save you time, and the reality of what we can do today? What are the actual changes someone’s going to see if they say, “Okay, I’m ready to take this step, I want to be found, I want to take advantage of this new technology”? What’s it really going to look like?

Willem: On the financial side — is that specifically what you’re asking, Floyd, or —

Floyd: We know it costs less than it has before, and we know, as Jay said earlier I think, that the price of not changing has never been higher. But day to day, how much of my day is going to get blown up if we decide to go forward with adopting better technology to handle our product data?

Willem: Yeah, well, just to talk a bit about the financial aspect — we do fixed bid now. We can be very precise in delivery, we know what’s coming up, and for that reason, unless the scope completely changes once you get into it, your price is what we tell you at the outset. That’s a huge benefit for folks.

Matt: I think the real challenge you might be talking about is: I have a day-to-day job I have to get done, and you’re going to throw a PIM implementation project on top of my plate. How is AI maybe helping alleviate some of those concerns?

Willem: Well, you need to talk to the business — there’s no getting around it. If it was just lift and shift, we wouldn’t need you. But we’re going to make changes, so the benefit is that you just ask people to show up for these detailed discussions. Those get recorded, the recordings end up in an AI delivery system that creates user stories, and then you need triage of those user stories. There’s far less back and forth, because things just move faster, more accurately. But there’s no getting away from wanting all the SMEs in the room really paying attention — because if they don’t, I’m trying to keep your rating, Floyd, but they’re just not thinking straight, because this will have a giant impact on them down the road. The challenge I see a lot is people want things online to be perfect, they want great PDF data sheets, they want a lot of things, but they always struggle to spend time creating those foundational capabilities.

Jay: Floyd, I think you may be getting only a subset of where people need therapy. If you said to people, “Talk to me about the therapy you need for maintaining this in Excel or a legacy system, talk to me about the therapy when your exec suite says, ‘No, you don’t have six months to launch this, you have six weeks, go’” — there are levels of pain people have internalized that are worth questioning. I really like what Willem said, that you need all the SMEs in the room at the time, because they’re the ones who can start to internalize this. I’ll make it concrete for our listeners — we surveyed about 65 customers and asked them to tell us about the after experience, what they’re seeing. More than two-thirds said they reduced time spent maintaining product data by more than 30%, and 75% said they’re launching products faster — on average, about 30% faster. So there’s always a step that’s now much shorter than it used to be, in doing the transition, but your ability to run faster once you’re on the other side of that transition is much higher.

Willem: Yeah, I like how inriver deals with additional channels that come up — for example, the way they can pull data out of the product repository and create filters that pull that data to the channel, rather than creating an additional push that needs configuration at a much higher level. That’s a great example of being able to scale in a much faster, easier way, with all the configuration being very much a UI configuration, rather than somebody like Pivotree having to come back.

Jay: Floyd, the other place I think might cause some of our users to lose their clean rating is when they suddenly find that manufacturing for a given part has moved between countries, or that the EU has imposed new compliance regulations if you’re selling there, and what information you need to maintain to handle that. So how are you setting your systems up to handle that degree of flexibility? That’s not only one of the benefits you get from the transition, it’s why it’s important to work with companies like Pivotree that have actually done this enough times to have the experience of knowing what you want to be thinking about as you’re future-proofing your solution.

Matt: I was just thinking about this valley of pain that happens in any platform implementation, no matter what it is — you start to get fatigued. You guys were talking about marathon running before we started, and the idea is there’s a point where you start forgetting what you’re running toward, and you start focusing on the pain. Maybe my analogy breaks down, maybe it doesn’t — but a good SI is going to remind you what the finish line looks like. It’s going to remind you, “This is why we’re doing this, yeah, that’s painful right now, but here’s why, and here’s ultimately the outcome you’ll benefit from in the long term.” Having that big-picture perspective, not getting lost in the weeds of the implementation, but continually being reminded of the transformation — not only to the team, but to the organization, the customer base, the market — that’s really important to keep the engine running and everybody excited and pushing forward.

Jay: I’ll keep the running analogy going — if you’re actually running a marathon, running Boston, you run for the Wellesley Scream Tunnel, you run for the firehouse, you run for the point where you can see the Citgo sign. A good SI is going to help you understand what these milestones are that you’ll celebrate, that take you through the pain. And look, let’s be honest, the pain has actually come way down, as Willem said at the beginning, because there are so many new tools that make it faster and easier to architect, implement, and adopt the end solution.

Willem: Yeah, people tell you smiling helps — if you smile, you feel less pain when you run. I think getting to the end, that middle portion where people start losing focus, is a problem. It often means bloat of requirements — people start thinking, “Oh, well, we could—” they start understanding the platform, they understand everything, all of a sudden light bulbs go off, and everybody wants to do everything. That’s where you really need strong management guidance and an understanding that if we do everything, we don’t get to the finish line. So let’s focus and get it right — a really important part of managing the project to the end goal.

Jay: Mm-hmm.

Matt: Yeah, I love that. We — and when I say “we,” I mean anybody in the technology space, whether you’re the company we’re helping implement or a software company — have a tendency to over-engineer, because we have everybody’s best interest in mind, we see what ideal looks like, we’re all idealists at heart, we want the right thing. It’s not about doing everything, it’s about doing the next right thing — roadmapping so you’re achieving the long-term vision, but also able to put something on the playing field as fast as possible to get the fastest return on that investment.

Jay: And Willem cited the ability to build prototypes up front, but these days a lot of those prototypes can be turned into implemented code, or a pilot that’s in production very quickly, and thanks to MCPs and other technologies, integrated well with the rest of the stack without nearly the pain it used to require.

Willem: Yeah, “prototyping” is actually the wrong word, Jay — I don’t know what to call it, but you’re right, it’s not a prototype. If you build on top of it and it actually becomes the final solution in the end, I’m not sure what we should call it, but it’s not a prototype.

Jay: I’ll assume our audience can assign whatever word to that they want. But it’s something that wasn’t easy to conceive of even two years ago —

Willem: Yeah.

Jay: — because it’s no longer just wireframes, it’s working code.

Willem: Yeah, exactly.

Matt: And the software — it’s like a rapid MVP, right? How quickly can you get to the minimally viable product and then continually iterate on top of it? Yeah, that’s incredible. I’m not going to make promises, but we’re talking days — this is the incredible thing about what we’re able to do in terms of standing up a platform, even a PIM, with all the complexities. We can get in there and get our hands dirty really quickly these days.

Jay: And that actually helps organizations evaluating how much they want this — it’s easier to see what the end result is going to look like, and start to say, “Okay, I can actually see the finish line,” without over-torturing our running metaphor here, a lot more easily than it used to be.

Floyd: I think speed has really been the theme of the show — it’s faster than ever to see the benefits of making this kind of investment, you get to market faster, you free up like 30% more time for the people who’ve been doing this manually over and over again. I think we’ve got maybe the fastest guests we’ve ever had on the podcast — I don’t know if you’re running any marathons in the near future you want to plug?

Jay: That would be Willem —

Willem: Jay did Boston, so he must be faster than me.

Floyd: I could tell — that was some real insider knowledge.

Jay: Well, I live not far from the course, so all of these are familiar landmarks.

Willem: Yeah — on that Newton Hill, Jay — but no, it’s a brutal course, and I’ve never qualified, so I’ve never run it.

Floyd: Wow.

Jay: You’re welcome to come run the Newton Hills anytime you want, Willem — they’re not far from me, and yes, they’re brutal, but they’re also fun. So Floyd, save us from ourselves.

Floyd: I’ll save you, I’ll save you, I’m swooping in — I’m not a runner, I’m sturdier than that, not built for speed. But I hope people run to our LinkedIn Live event, if I may beat that metaphor one more time — this episode coming out today, Wednesday, September 2nd, later this afternoon, we’re doing it again, plus Jay’s organization, inriver, has done some really interesting research about product data maturity. Just a little preview: you surveyed more than 400 manufacturers and distributors, something like a third of them said pretty confidently they were ready for agentic commerce, and when you dug into it, zero of them actually were. So we’re going to talk about what that meant, and what other manufacturers and distributors can take away from that. Look us up on LinkedIn — we’ve got a link on the Data Versus Commerce page, on Pivotree’s page, and inriver’s page, and we hope to see everybody there. If you’re listening to this after September 2nd, go to the same places and you can find the recording. Jay, Willem, it’s been awesome to have you on for the series — where can people find you if they want to talk more about PIM?

Jay: People are welcome to reach out to jay.roxe@inriver.com, or just find me on LinkedIn.

Willem: Yeah, same here — LinkedIn works, or willem@pivotree.com.

Floyd: Nice — oh, you left your last name out, that was a kindness to our listeners.

Willem: Yeah, it’s easier that way.

Floyd: Thanks so much, see you next time.

Matt: Bye, guys.

Willem: Yes, bye.

Matt: Thanks for tuning in to this episode of Data Versus Commerce. New episodes drop weekly. So if you’re responsible for any part of how products get from a database to a doorstep, subscribe now on Apple, Spotify, or wherever you listen.