This week, Andy & Brian talk about whether large, complex, global enterprises should centralize their digital operations, and how to think about that decision.

They dig into 4 areas:

1) To what extent the market is dominated by local preferences and behaviors

2) What are the service components of the business?

3) What are the fulfillment requirements?

4) Where is talent located?

— 

Brian Beck: Welcome to Friday 15 with Master B2B. My name is Brian Beck here with my partner in crime, Andy Hoar. I’m broadcasting here from Amazon in Washington – that’s Seattle. Overcast weather as usual, Andy, but overcast in the summer is kind of unusual for Washington because typically in Seattle it’s nice and sunny. That lasts about four weeks.

Andy Hoar: That’s why people move there, because there’s never any rain.

Brian: That’s right. Never any rain. Well, folks, welcome to Friday 15. We’ve got some pretty interesting and exciting breaking news today. Let’s get to that. Andy, did you see this? Last week, Google’s search monopoly ruling, they were ruled against by the FTC, I believe, right? That said, hey, you’re a monopoly. Whoa. What’s the deal here, Andy? Talk to us about this and what the implications are.

Andy: It’s just eerie how similar this is to what happened to Microsoft back in the 1990s. And so anybody who’s under the age of 30 back in the 1990s, Microsoft was embedding the Internet Explorer search browser, which doesn’t even exist anymore, on the desktops of all these PC manufacturers like Compaq Computer, which also doesn’t exist anymore. And they were paying money to do that. So Compaq would get paid billions of dollars to put the IE browser on default on the desktop. Well, fast forward 30 years and the same thing is happening all over again where now it’s Google putting its search engine, well paying to have its search engine embedded in browsers that didn’t exist 10 15 years ago like Firefox and Safari but they’re paying Apple for example to be embedded in Apple’s browser. So these tie-ins are kind of the centerpiece of this and the FTC found that paying money to be the default search engine now on these browsers is creating a monopoly. Now, the interesting part about this is it’s hard to prove that there’s been any harm here because the court had a finding of fact that Google search engine was the premier search engine in the space. On the other hand, they said, we don’t have the ability to easily click off, even though you can just go to Bing.com In the browser. But apparently, even everybody agrees there’s something called choice friction. And Google’s in-house lawyers said this, too, in emails, which was really stupid. But they said that’s asking people too much. It’s asking too much for people to just go up to the browser and type in Bing.

Brian: That’s ridiculous. This is an indictment on the laziness of U.S. consumers, I think.

Andy: That’s exactly what it is. And so the question is, how are they going to relieve this? I think you said in the next slide, the Wall Street Journal had an interesting story about this, saying that Google is kind of a strange monopolist. And they pointed out that the judge found that Google would achieve market dominance because, and I quote, “it had the highest quality search engine, which has earned Google the trust of hundreds of millions of daily users.” So they got the best mousetrap out there by the finding of fact. The most interesting part here is that some of the key testimony was given by Microsoft, which owns Bing. And Satya Nadalla said, oh, this nightmare scenario where, you know, Google could use their search history to train an AI bot. And it’s like, oh, really? Like the one Microsoft is an investor in called ChatGPT, which just launched its own search engine called SearchGPT? So this is two three trillion dollar companies kind of fighting it out over who and but the reality is and you and I talked about this beforehand that they can complain all they want but by the time any of this actually comes to pass this will be in the ash heap of history. Five years from now Google’s search dominance may be nowhere because we might be using ChatGPT for a lot of this stuff. And Google, for example, may continue to fumble their AI rollouts. And Microsoft may be the dominant search property on the planet.

Brian: Could be. I still get back to the harm thing. What is the FTC doing pursuing that? If there’s no real… consumer harm here the consumers – maybe that that friction thing is real, I don’t know. It’s sort of like “hey are we just sort of stupid consumers out here and we don’t know how to switch?” There’s actual choice and frankly if someone brought out a better search engine wouldn’t it work to the advantage of the browser companies to build that in?

Andy: There’s a reason why Apple, which could create a search engine, could buy a search engine, uses Google search engine. I’m not an apologist for Google. I’ve got issues with how they do things too. But the harm here is really a bank shot to say that because Google search engine is so good and companies are paid to embed it, that people who want to advertise and buy keywords have fewer opportunities because Google is kind of the only game in town. If somebody has a better search engine that would eliminate the problem. And that’s probably what’s going to happen here until these GPTs are going to come along. I don’t know about you, but I don’t use Google search as much as I use chatGPT. So maybe we’re already along this path.

Brian: Maybe. I do think letting the market play out here and someone comes up with a better search engine, people are going to gravitate towards it, including the browsers. Google now has a legal ruling that they’re the best search engine so that’s kind of interesting too that jumped out at me. Let’s get into our topic. This is a fascinating one – we could do a whole Friday 15 just on that. But our topic today is should the e-commerce function for large, global B2B companies with diverse lines of business be centralized or decentralized? And Andy we’re talking here about companies that might have operations in 20 countries and have 30 brands they’re operating with, lots of diverse business units. So this is really we’re speaking to large global manufacturers, distributors, people like that here. And it’s a question that you and I hear about sort of all the time. Why don’t you take us through kind of what you know, what does it mean to kind of what’s the issue here?

Andy: There really are two schools of thought when it comes to this org design stuff. And this matters, by the way, because the way you design your team will actually produce certain outcomes. It’s just a fact. If you have a very centralized team, you’re going to get a certain set of outcomes that flows like water does down a river, right? And so it does matter how you design this. And so again, the two schools of thought, centralized versus decentralized. This is a poor graphic to show it, but one is where you put everything in place and it all gets distributed out in elegantly or inelegantly. And then the decentralized is where you forward deploy a lot of these capabilities. Now, we have been talking about this one for a while. You and I have had lots of conversations. And in fact, a year ago at our Mindshare Summit, we asked people to tell us. We handed out paper and we said, design your org chart. And then we hung them on the wall. And I think we had something like 96 different versions of them. We put them on the wall and it was fascinating to see people walk around and point at them and say, Oh yeah, this is a good idea. Or we hadn’t considered that or what have you. So we’ve been talking about this one for a long time. And so it’s not a new concept, but what is interesting is we like to think about this. You and I have talked about this quite a bit. What are the factors that you need to consider when you’re making a decision about whether to centralize or decentralize? And this is not an exhaustive list, but we think these are kind of some of the biggies. So one big one is to what extent is your business is really dominated by local customer preferences and behaviors – currency, language, that kind of thing. If you’re operating a multi-brand business and the brands really are unique to a certain country- like I remember when I went to Japan many years ago it’s sort of funny, there was a soft drink there called it’s spelled c-a-l-p-i-s and when you say it it’s pronounced “cow piss” and you’re like “that’s great.” But say Calpis, that’s the name of the drink. But in Japan, people don’t think about it that way. So obviously if you put that in front of a group of Americans to market that, they’d change the name.

Brian: It’s like, Andy, it’s like the Nova in, I guess it was Chevy or whoever released it in South America. Well, Nova is a car, Nova. And Nova means don’t go. It doesn’t go. Exactly. So it’s the same thing, right? But I think this applies not only to geographies at this point, but it also applies to business units, even in the same geography. So if you think about a company that may have different types of products or different brands or different product types, or even serving different customer bases, To me, that’s another flavor of decentralized. You need to give, I think, some capabilities to those people to react to their customer, which might be different than the other lines of business the business is in. So I think this cuts across geographic and also business unit lines.

Andy: The second thing is a service component. And what I mean by that is, does he have to install this stuff? If you’re buying windows and you have to install the windows, obviously the installation is a part of the process versus if it’s a widget you ship and there’s no real local dimension of the fact that it’s delivered and it shows up. That proportionality matters as well. The fulfillment – how difficult is it to fulfill? Can it be shipped in? But perhaps the most important factor here is really the location of the talent. If everybody wants to live in a certain country and that’s where the talent is, then you’re gonna have to probably forward deploy that particular function and not force them to move to another city where they don’t wanna live.

Brian: It’s also where the talent is available. And then when you think about a centralized function, one of the key arguments is really about developing a center of excellence and expertise concentrated at a corporate level that can then be shared. Think about digital marketing or e-commerce platform development and experience development elements, things like that. That is really critical or could be a big advantage to a company. Whereas if it’s spread all over the place and you’re hiring in markets where there may not be as much sophistication or talent, that becomes more challenging. And that’s a big deal. I agree with you. I think that’s the number one factor in my book.

Andy: I think you pointed this out, too, when you were talking about your post on LinkedIn and you talked about if you centralize, say, the tech infrastructure, then there’s some economies of scale associated with that. If you’ve got 25 different versions of platforms all around the world, you can’t really benefit from a bundled discount. Why don’t we go through each one of these and give our thoughts on it. Which of these needs to be centralized versus probably more effectively decentralized for a global B2B manufacturer or distributor?

Brian: For those of you listening on the podcast, here is a list of key functions affected. We have six of them laid out here. So let’s talk through them. So data, consolidation and normalization of data, should that be handled at a corporate level or down, distributed down? And I think I’ve seen this firsthand with companies I’ve worked with in the past. And for example, I remember years ago working with Epson, which is a big printer manufacturer, and they were in 18 countries across the Americas. And this was a big deal for them. They had to allow for the variance of product data at individual country levels because there were different components, features, different requirements in that country for power, for example. Powering up a printer, right? The power configurations are sometimes different in different countries because the electrical grid is different. So I think this can be centralized, for example, but it has to allow for some, at the edge, it has to allow for some customization, product data customization. So that’s when I think you need some sort of, a bit of a federated model.

Andy: The counterargument is that data, if it is in a central location, can be harnessed more effectively. You can have data scientists in 14 different countries all leveraging best practices versus a centralized model where you can have 14 data scientists working together in one location. I think because we’re earlier on, I would tend to lean towards centralizing the data until it’s more advanced, where perhaps you could decentralize it at some point. But you need to centralize the talent, I think, and the understanding.

Brian: I would agree with that. Tech infrastructure, I think, is one that certainly has benefits if it’s shared. Again, using these distributed models where companies may acquire other brands and they have legacy systems or different markets. Some markets might have e-commerce, others don’t. You’ve got individual instances. It’s kind of a mess. And if you don’t centralize tech, I think, in my opinion, you end up with just a hodgepodge, a lot of technical debt, a lot of difficulty in maintaining systems, and then ultimately, you don’t have the same leverage with vendors to negotiate a great contract because you’re negotiating one-offs in all these different markets and countries and business units. This one I think has to be centralized.

Andy: A decentralized version of tech infrastructure often becomes a customized version of tech infrastructure, which is impossible to upgrade at some point, and becomes costly to maintain. So this one pretty clearly lies in the other direction in terms of centralization.

Brian: I would say marketing though, Andy, the next one on our list here, this one I think has to have some localization, whether that’s business unit or geographic. And it’s because you’re speaking to a customer group here that may have different needs that vary by country, even the language. And so marketing, I think, while you can have some centralized expertise around marketing analytics, around SEO, paid search, things like that, when you get down to the language and the distinctions between the different markets, I think you have to have some distribution here, in my opinion.

Andy: Isn’t it interesting? You can take the data and the tech infrastructure out of marketing and centralize that, a la our point of a minute ago, right? The other part was sales. I think the sales data and the sales tech infrastructure can be centralized, but the actual people have to obviously be on the ground. Customer support. Boy, that’s an age old question. Do they need to be in market or not? I mean, we’ve experimented with multiple models here in our lifetimes. I remember offshore, onshore, near-shore. Who knows? But i don’t have a good answer on this one I think that every company is different.

Brian: I think there is some benefit to centralizing this at least to a large degree. And then operations and fulfillment at the end of the day you’re going to need localized fulfillment here. I used to work with a company called Teleflora back in my consumer days. And they would have localized fulfillment, but they’d also ship from all these florist entities all over the United States. But they’d also ship some of the common goods, meaning the stuff that was shared across all of these florists from a single DC around the United States. So I think we’re going to end up here with a hybrid. So let’s talk through what some of them said. I mean, here’s a quote from Ian Heller. I’ll read this one quickly, then Andy will get your comments. Ian’s the founder of Distribution Strategy Group. He was formerly at HD Supply. And he said, “There’s a careful balance here, and it’s not easy to hit. The optimal solution is the messy one, which is often the case, a hybrid that is unique to the company based on its talent profile, market, and business model.” Yeah, amen. I agree with that, Ian. What are your thoughts?

Andy: I think you nailed it. That’s exactly what we’ve been talking about. Decentralized versus decentralized really depends on different functions and overall what the goals of the company are. But you can’t just take one model and shove it on top of it and say, everybody should decentralize or everybody should decentralize. It really is based on your business model and basically your go-to-market strategy and what your customers prefer. Because to the point of customer support, it could be that having people in the Philippines speak to people in the United States is just fine. But it could be that having people in the Philippines speak to people in the United States of America doesn’t work. And so it’s specific to the company itself.

Brian: We got a lot of comments on LinkedIn when we posted this, Andy. Here’s another one from Tim Peterson, President CEO of Spear Digital. And he said, “Central at the very top with country area brand leaders, whichever is needed around the organization.” That’s his recommendation or his thought. Then you get both local expertise, control and learning with cost savings globally. Again, kind of a hybrid thing. What are your thoughts?

Andy: One area we didn’t talk about was finance. And, you know, that’s a question that’s often asked, too, is did you centralize the financial component or decentralize it? I think for the most part, companies centralize that because they’ll be able to track it. But I’ve often heard people say, how is money any different from data? Data and money are kind of like the same concept and they both flow. There’s a good reason to centralize both of them. So you wouldn’t decentralize the finance of the organization. Why would you decentralize the data component, right? I think there is an analogy between the two. And like you pointed out, cost savings, you can’t really do that at the local level if you don’t really have a good sense for how to normalize it. I think the money should be centralized. And I think when you do, you do tend to save some money, at least in terms of licenses and implementation, et cetera.

Brian: We just got a comment in on Llinkedin – thank you for that comment. “I believe the key factor here is how customers behave and what they expect from the company’s location while centralizing data governance and fin ops can be great if customers are used to a different workflow it might lead to more complications than benefits.” Any reaction to that Andy?

Andy: If companies have workflow, it’s where the rubber meets the road. In fact, when you think about implementations, how long does it really take to move data around? You know, not very long. All of the complication with an implementation of a platform, for example, is having it interface with people and their existing processes. And that’s what the person just pointed out is that they have established workflows. This could complicate those workflows. Well, what if the workflows are inefficient to begin with?

Brian: Right, that’s a good point. And thank you for your comment there from, from LinkedIn. So wow. And it’s like, I was getting all kinds of comments. I can’t share all of them, but Robert, thank you for your comment as well. So well, let’s, let’s just share it real quick because it’s interesting. Robert Calvert, thank you. “Isn’t there a broader centralization topic if the e-business group has a single centralized function versus embedding business functions with a matrix overhead?” I think there’s a lot of back and forth on this. And I think where we’re landing and the community said it when we went out to Linkedin we asked this question: Should e-commerce functions for large global b2b companies with diverse lines of business be centralized or decentralized? When asked 64% of them said centralized – that’s two-thirds. I tend to lean to that, but there’s got to be, in my opinion, some level of extension out to the business unit, to the local market.

Andy: I also wonder if this is a self-fulfilling prophecy. We’re asked about large companies. multinational global B2B companies, which already have this centralized framework, right? They start with that. When you ask them, well, is this the right version? They’re like, well, that’s the one we’re using. Therefore, it’s probably the right one. So it becomes a confirmation bias. That said, I do think largely speaking, it’s better to centralize versus decentralize, if for no other reason, just for control. You don’t know what’s going on. Yes, you can have little experiments taking place and there’s a value in doing that. But you could also have 25 different versions of things, and it becomes a mess.

Brian: Well, I think talent’s the overriding driver here in my experience, and working with a lot of different companies that deal with this issue.