Skip to main content


Building the 5G ecosystem

Podcast episode 36

5G requires additional CAPEX investments, and it’s arriving at a time when telecom companies are feeling the pinch. But to Salesforce’s Steve Mannel, the real ROI for 5G won’t come from consumers -- it will come from building an enterprise ecosystem.

Below is a transcript of this conversation. Some parts have been edited for clarity.

Michael Hainsworth: We've all heard it. 5G isn't just another generation. It's expensive to deploy and operate. And it's arriving at a time when telecom companies and other communications service providers are having difficulty stemming the slide in average revenue per user. But to Salesforce's Steve Mannel, The real return on capital for 5G will take place on the enterprise side, but only if a CSP can shed its 4G thinking.

Steve Mannel: The predicted return on capital that we've grown accustomed to with previous generations may not largely be coming from the consumer side. Consumer growth has, in many markets, been saturated as it relates to population growth, and the number of handsets that people are willing to carry. And the path to revenue with value added services, whether it be a premium version of connectivity or other yet unknown devices or applications, yeah, the predicted return on capital from the consumer side is flat to marginal growth. So because of the expense of 5G, right, we've got all this spectrum at the high, middle and low end, all these new radios and towers and backhaul, right? There's a pretty significant expense related to 5G without that formerly predictable return on capital. So a lot of folks are looking to the enterprise as fertile ground. Not only because it's the only place to go theoretically, or one of the few places to go, but because of the new sets of capabilities that are enabled by network as a service writ large. And 5G and MEC (mobile edge computing). So the opportunity appears to be poised in the B2B or B2B2X space. And that's pretty much consensus at this point, right? I can't seem to find an analyst that doesn't seem to lean into those assumptions and the pending value propositions.

MH: Let's talk about fertile ground. We have devices, applications, managed services and APIs that drive data co-mingling together in this new way. That must have forced Salesforce to evolve from just a CRM company.

SM: Well, yeah. So, two or three years ago, if you had asked me what Salesforce's position or point of view on 5G would be, I probably would have said something like, well, CSPs are going to sell 5G differently, right? So that would impact marketing sales and service, kind of our core lines of business. But in some respects, we would have been secondary or even tertiary to the larger conversation, right. But our portfolio has grown and changed over that interim time period. So we were forced to look at the market and our portfolio through a completely different lens. And that led us recently to commission a study from a company called STL Partners which we've entitled Taking 5G Beyond Connectivity: Growing the B2B2X Business. This paper was recently released. And the thesis there was if this 5G business is starting to look like a platform with an ecosystem built around that platform and eventually a marketplace, what would Salesforce be able to bring to bear in that context? And it forced us to rethink our point of view completely. We tend to look at our existing opportunities and customer stories as the basis for not some sort of fictional point of view, but something based in reality, and we saw a completely different outcome.

Yes, we do think that enterprises are going to market, sell, and service differently with the biggest material difference there being probably that we're going to move away from a wholesale large enterprise SME consumer type of model where we've been selling circuits and lines and voice and security products into something that looks a lot more verticalized.

So that would have been sort of the big punchline, if you will, of our point of view a few years ago. But with the addition of some capabilities in our portfolio, not to mention some critical acquisitions like Tableau, Vlocity (now called our Industry Cloud), as well as, and perhaps most importantly, MuleSoft, which is a market leader for API networks, is the sort of term of art that we use. We have the potential here to now not only affect the way things are marketed, sold, and serviced, but also encapsulate the way network as a service is packaged, right, as an API network around which you can build developer community something we know quite a bit about, with 15 or 16 years of experience there. And then eventually building out catalog driven marketplaces in, in light of this sort of B2B2X construct which seems to be the predicted monetization models.

So, we went from being a secondary at best sort of solutions provider that would address this transformation of the way things are bought and sold into now sort of hopefully primary in the conversation as it relates to monetization of these new business models, enabled and powered by network as a service at large and 5G in particular.

We opened up the platform. We attracted an ecosystem. And we've turned that ecosystem into something much larger than ourselves.

MH: If Salesforce was forced to evolve into a platform with an ecosystem and a marketplace, and it all evolved out of that, what is an ecosystem for a CSP now under 5G?

SM: Software as a service is not platform as a service, is not infrastructure as a service, is not network as a service. So I think they're all going to be some nuances and differences as it relates to the ‘X as a service’ models. But I think in some respects, there are a lot of parallels and lessons learned and best practices.

When I got here in 2009, I felt as though our language, our verbiage, around being a platform may have been out of our skis a little bit. We were an award-winning applications company that I think aspired to be a platform. We had all the earmarks of a platform, right? We, if you think of that software as a service and platform as a service, we started to take this 80/20 rule, right? It's maybe not for all workloads or applications, but most, and then inside of that most category you could do most of what you want to do through clicks, not code. That's sort of our stock and trade, right? So when you come to the realization then that 80/20 rule and 80/20 clicks not code could apply to other people's business logic, you get the idea that you can open up the platform and let other people in to develop applications around business functions that remain.

MH: Right, this is a huge cultural shift that we're talking about here.

SM: It is, but there are certainly, let's call them leading indicators, about why you may want to do that, right? So for instance, we work with IDC on a paper called the Salesforce Economy. And I think it helps illustrate a lot of the questions that I often get about Salesforce. A lot of the questions I get are, you're only at, let's say in this case, a $20 billion company, how do you explain your market cap, right? You're only a tier one enterprise software company. How do you explain that you can get 190,000 people to San Francisco every year for Dreamforce, right? How do you explain that these global systems integrators have built multi-billion dollar practices around your platform? And the only way I can explain it is that we saw the potential of the platform. We opened up the platform. We attracted an ecosystem. And we've turned that ecosystem into something much larger than ourselves. It's the network effect, it's a blossom, right?

So we're this tiny little boat that's pulling this sort of massive wake behind us, that's created an ecosystem. And I think there are many, many parallels that we can draw, whether it's B2C, or, in the technology space especially, where those things should become obvious. It doesn't mitigate the fear and the need potentially for this leap of faith that I think we're talking about here. But there's a lot of lessons learned and best practices that can be gleaned from early successes, for sure.

MH: As the originators of SaaS, you talk about southbound systems and northbound systems. A CSP is interested in growth of ARPU, not just stemming its slide. And you point out it's the northbound systems that provide this Greenfield opportunity. Let's define what north and southbound means and why the real growth opportunity comes from the northbound side of the equation.

SM: Sure. So, I typically use the terms northbound and southbound in the context of systems and APIs around those systems. So your southbound systems in a CSP’s case would be your network and the products you build around that network. You've got your operational support systems for service assurance. You've got your revenue management systems, your BSS, as well as your traditional enterprise IT applications, human capital, management, ERP, et cetera, et cetera. So in a way, we're in a new world as it relates to getting your enterprise house in order, with cleaning up your API and your integration strategy. That's a little messy, because you're going to run into tens of years of integration tools, strategies, some of which are homegrown, open source, it's kind of a mess.

So the more Greenfield opportunities, as I see them, are on the northbound side. So this is where we start to get into the lessons learned, right? So Salesforce is a platform, okay, great. Do you have APIs on that platform? Yes, we do. Okay. Well, how do you build an attractive developer ecosystem around those APIs? Well, that's not something we sort of birthed whole year one right? I've had many, many conversations with the folks who built our app exchange over the years, some of whom I've become very, very close with, and it was a work in progress. And what we ended up with was a very robust set of tools, or we largely call partner communities or partner experiences. And we've married that up with some tools from MuleSoft around our API boundary layer, right? So not all of these applications that these APIs that are going to be built are going to be exposed natively, right? They're going to be exposed through a developer community with boundary systems for security purposes, management, all those sort of typical things that you'd expect. So that is where I think there's a lot of fertile ground.

What we've seen from the CSPs historically is they'll build a few dozen APIs. Maybe they have what I would call a partnering program, which has a select number of people who can get in to access those APIs. And it's very tightly controlled. It's kind of a closed partner ecosystem. It's kind of a controlled process and it doesn't really allow for serendipity. It doesn't really allow for discovery. So that's sort of what I mean by us bringing a lot of our tools, which can be white labeled, which would help attract and retain a developer community with all the bells and whistles that need to be in there, right? And then that can evolve into more of a marketplace.

MH: Then when we talk about where CSPs can add value to enterprise, we should talk about their new role in defining a digital economy. You cite Apple as a perfect example of a digital economy that CSPs should model.

SM: In a way, yes. Apple put out their iOS platform. And if you recall, they brought, I think, somewhere in the neighborhood of $400 million into the developer ecosystem to sort of stoke the App Store, so that when the App Store came about it wasn't a barren landscape, right. You wanted some richness there. And I think that could be, again, a parallel here. A lot of us as enterprises, right? Not just not Silicon Valley, but enterprises, have been working around IP transport for a long time. It's just something that we couldn't rely on. Its burstability, right? We couldn't turn it up and turn it down. We couldn't define its location. We couldn't cash capabilities at the edge of the network. So we've been programming around the network and over the network for a very, very long time. And I think the opportunity here is to understand, identify IoT devices, applications, managed services providers who understand how to leverage the network and help illustrate to the rest of the world these new use cases. So that's really the opportunity that I see. And I see a couple of operators doing a very good job of that already. We've seen some of the investments gone, unfortunately, being a little less, let's call them modest, right? If you're spending a few billion dollars on your network, but you're only willing to spend a few hundred thousand dollars or a few million dollars on stoking your platform ecosystem, it's disproportionate and perhaps not conducive to the end result.

Now we're in a dynamic situation where we can potentially charge more for things.

MH: Right. It sort of feels like cheaping out on the architect when building a skyscraper.

SM: Right, exactly.

MH: So then do you find that there's still a great deal of skepticism about what it means to build out a network as a service, that skyscraper?

SM: Oh, without question. It's scary, right? In some respects, some of the feedback that we've gotten in the course of interviewing 23 different chief strategy officers and heads of network in the operator space for this recent paper. And a lot of them are intellectually there in terms of the leap of faith but they're also worried about further commoditization. I think we understand and appreciate those concerns, right? We've seen this sort of cost per dollar of transport continue to decline. And the idea, I think, in some corners is that if you expose the network, it can become more commoditized. But I think we're baking in some capabilities into the network for turning this into a marketplace, and that we could actually turn that network and those capabilities into premium services if we stoke the ecosystem so much that there is demand across the available assets, right? So, it's like any other marketplace. I use the word deliberately here. So, to the extent that we can fill the marketplace with demand, that's going to meet or exceed the capabilities of the network, the supply side. Now we're in a dynamic situation where we can potentially charge more for things. So, that is a part of the intellectual leap that I think people are struggling with because they don't see a clear path to that.

MH: You've been quoted as saying you advise CSPs to leverage localism. What does that mean?

SM: Localism is, I think, a good way of contrasting what happened with iOS and Android. With iOS and Android, we built a global platform or two, which we needed, in order for, application developers looking to take advantage of smartphones, right? And essentially, they achieved scale. In this context, network as a service, and particularly things like MEC. The advantage isn't scale so much as it is localism, right? Having local points of presence. Having capability at the edge or near the edge. So owning spectrum, owning radios, owning edge caching or edge computing capabilities is the advantage here. And the challenge will be to expose that to developers and educate them how to explore those capabilities. That's what I mean by localism.

MH: So how can platforms and ecosystems provide the right environment to aggregate value in the digital era?

SM: Oh, that's a really good question, and one that I think we see a couple of operators doing a good job with. So the TM forum and STL, they define this as a continuum, right? If connectivity is the core business along with communications and network security, networking service enablement is the sort of next chapter, right? Service enablement in some respects is really enabled by network as a service. There's a certain third sort of model, which is service orchestration.

The difference between the second and the third model is that I could be the application provider that's licensing the networking capabilities. And I pass that at the top line along to my customer. And the customer doesn't even know that it's the CSP's capabilities underneath, right? That's not mutually exclusive to the orchestrator model or the third model that I was referring to earlier. But the third model sort of leans into this notion that the network is what it is in terms of a commodity. Now there might be some bursting up and down on that commodity, but there's a value around that network. There's devices, there's applications, there's managed services, there's professional services. And if we can expand the aperture of what it is we do as a CSP, if we can bring a curated set of solutions into verticals that allow me to take the top line revenue and be that service orchestrator, right. You've heard the expression “Software is eating the world.” Well, this is an example of leaning into some of those spaces that allow them to take that software-driven application business value model and deliver that, right? While fully exploiting the underlying capabilities in the network in a way that someone who's architected around the network may not be able to do. So, that's sort of what I'm referring to there is this notion that we can potentially expand the pie just by being better at exposing the network so that the service enablement business is optimized. But not mutually exclusive, we have this other opportunity to curate a set of solutions, right? Probably for verticals. And take the top line revenue as a result and be that service or solution orchestrator.

MH: Well, in Canada, Telus is an example of a CSP that's pivoted to a managed services organization. They've got the healthcare unit, the business outsourcing division. Is this orchestrator model the future of CSPs?

SM: Maybe for some, right. One of the conclusions that we drew from the paper was that not all markets are created equal, and challenges are not incumbents. So, we're not being prescriptive about what the model should be. What we're saying is that there's gonna be multiple models and that there are lessons learned and best practices along the continuum of creating those multiple models. Now we've seen some really good early examples, Like Tellus, or Reliance has an IOT business unit called Unlimit. And while I'm not in the Indian market everything I read about it sort of indicates that it's very much an open model where they have kind of a, they give the end customer choice as to the devices and applications and analytics, and all the other things that go into the creation of the solution. So they're not being, they're not being overly prescriptive about the the solution that they're orchestrating, right? They're letting it happen somewhat organically. And those are parallels that we were very familiar with, right. On the Salesforce app exchange, of the 5,000 applications in the app exchange, there's a lot of overlap. There are competitors that essentially do the same thing and we're not here to be prescriptive to tell the customer who the best is, right? Because while vendor X might be the best for a company A, vendor Y might be the best for company B. So we're not here to be overly prescriptive. So that's a sort of long answer to your question, but I think there are some early indicators and some early examples in the market that suggest that it could work for some operators in some markets. For sure.

MH: So then what would you say is the unique opportunity for CSPs in this digital economy?

SM: To better understand how to exploit conductivity, right? Right. If conductivity has been programmed around, right, the opportunities is to educate particularly the application companies where theoretically most of the value is, how best to exploit the network, right? I think we've had a long history of building technology for technology's sake in the communications industry. And a lot of that has been proprietary. A lot of what's different about now is that because most of these are being exposed, as rest APIs, something that most developers are familiar with, I'm not talking about Parley X here, right, from the mid 2000s, I'm talking about Rest, right? And I'm talking about methods and patterns that are going to be familiar with developers. They only need to be educated how to exploit them, right? And that's something that I think your average development community doesn't understand how to do. So, some of the early examples, you see the hyperscalers partnering with the large CSPs around their MEK strategies in an attempt to stoke that. But going back to your earlier question about scaring the daylights out of CSPs, in some respects that kind of feels like the Android iOS chapter all over again, right? You're inviting hyperscalers and their developers to exploit the network. And the question is where does the value go in that equation, right? And again, it's not mutually exclusive. The idea is to build a platform so you can attract an ecosystem and exploit the assets that you've got to their fullest potential to potentially create contention. At which point you can create a premium market for that contention. The opportunity to create that top line value is not mutually exclusive. And we've seen some recent examples of that as well, with the Edge Private Cloud announcements with the likes of AT&T and IBM, for example. So I don't think, we're seeing a lot of bets being placed and chips being put around the table. There are no single obvious bets. I think there's a lot of little bets.

MH: If there's one takeaway for the listener, is it about expertise or localism?

SM: Localism is your value. Teaching the developer community how to exploit the network is the challenge, externally. The internal challenge is not technology, right? It's people and process, it's culture. I sort of always summarize people and processes. It's culture, right? The technology, we're largely taken care of as an industry, right? The software virtualization of the network is largely driven by software companies or equipment companies that are learning how to become software companies. It's also being driven by telecom. Infra Project, MEF, 3GTP, MTM Forum, right? So a lot of that technology, I have faith in the technology. I have less face and faith in the culture. And I think there's a lot of people who will not survive the transition into the new culture. There's kind of, it has to be a lot of hiring from other parts of the ecosystem to glean that expertise. And then of course, you're going to have to train. So the culture is probably the biggest thing that needs to change. Having an open mind, a willingness to learn what it takes to build these things. And the arc of investment, It's a multi-year arc, right? Not unlike a big capital project on the network. So, a lot of what I'm advocating right now is rather than waiting for these chapters to unfold, network out, it's time to start building up the cultural aspects that's gonna help build the platform, that ecosystem in the marketplace now in parallel, otherwise there's going to be a missed opportunity.

<<Go to previous episode  Go to next episode>>