Jay McBain on How Microsoft’s AI Billing Passes Risk and Liability to MSPs

[00:00:12] For 30 years, the MSP business model has been built on one thing, predictability. You know your costs, you know your revenue, you set a flat monthly fee, and the math works. The entire recurring revenue stack, PSA, RMM, backup, security, was engineered to support that model. That foundation might be shifting. Microsoft's new E7 license layers an AI consumption charge on top of the per-seat fee.

[00:00:41] Every vendor with an AI product is watching that move and considering the same. And when machine speed agents are doing the consuming, not humans, the variable component doesn't scale like cloud storage does. It scales like leaving the faucet on with no shutoff valve. I've been tracking this on The Daily Show for the past several weeks. I've stress-tested it with MSP operators. I believe this is the most under-reported structural risk in the channel right now.

[00:01:12] Jay McBain has the vendor-side data. Today, we find out if my read is right or where I'm off. That's the tension that brought us here today. And Jay McBain is exactly the right person to pressure test it. I want to know where the vendor data confirms this trajectory and where my operator-

[00:01:29] level read may be over-correcting. Welcome to the Business Effect Lounge, live where I make mistakes in real time. This is where we break down what's changing in the IT services market and what it means for providers and vendors. And what to actually do about it. If you're running or supporting an MSP, this is about making sense of the environment you're operating in, not just the headlines. Now to make this conversation possible, a message from our sponsor.

[00:02:25] We'll see you next time. We'll see you next time. We'll see you next time. Jay McBain

[00:02:54] Jay McBain Jay McBain Jay McBain Jay McBain Jay McBain Jay McBain Jay McBain Jay McBain Jay McBain

[00:03:26] Jay McBain Jay McBain Jay McBain Jay McBain Jay McBain Jay McBain

[00:03:47] And there's a lot of justification around AI, but as you've put, the platform giants want partners that can implement, integrate, and manage, not just move product. Resell alone isn't. But then we layer on this new pricing model change. And I think E7 is definitely leading edge, maybe even as far as bleeding edge, that base per seat plus an AI consumption charge on top. That's usage-based billing layered on top of fixed seat.

[00:04:16] I look at this and sort of say, Jay, you've been tracking how partners are restructuring that. Do you think the move towards consumption-based pricing is going to be an industry-wide trend? Is Microsoft running an experiment? Is everyone watching? What's your take? Yeah, it's actually the move towards micro-consumption. So if you dial back a long way, this has been the fastest-growing industry for 56 years. So we've got ourselves to $6.07 trillion.

[00:04:45] This industry is going to double in size in the next three to five years. There's a $7 trillion build-out on AI happening, and we're just starting to understand the economics of it now. So the mainframe era, you know, the 60s, 70s on a lease type of economics, we moved into client-server through the 80s and 90s, the PC era, which was really a transaction model, that traditional billing.

[00:05:10] The cloud in about 1999, that's at the start of managed services as well, started introducing subscriptions to customers, where for a set amount per month, that's the predictability that you talked about. But a few years later, you know, Jeff Bezos had a crazy idea of renting out some extra capacity on their non-busiest days at Amazon, which today is a $600 billion consumption economy.

[00:05:37] But customers and partners to this moment have been somewhat insulated from that consumption. Customers at a big level have bought ahead credits that they consume. And in most cases, partners and a lot of the SaaS and other products end up consuming that on a variable basis, but then still charging a subscription. So here we are in, you know, 2026.

[00:06:02] Two-thirds of our industry is now subscription and consumption-based and quickly turning into micro-consumption. You're starting to hear magazine ads talk about customers getting killed with tokens and trying to manage these new budgets. And these budgets are going up by, you know, 35% while the company itself isn't growing and layoffs and all the other things that are coming from it. But it's not just Microsoft.

[00:06:27] There are hundreds of companies now that have introduced these micro-consumption models, which is $2 per conversation for a CCAS or UCAS type of solution, $1 per photo edit, or, you know, many, many other ideas all the way down. So how do we deal in this world of traditional billing, subscription billing, consumption, and now these $1 per thing micro-consumption? And that's your thesis is that this is unplanned for.

[00:06:57] And when you run kind of a fixed, you know, charge per month on a predictability basis, there's no room in there for variability. Well, and you didn't even include speed, right? This is at machine speed. The one thing that we could talk about, like we can compare, try and compare this against, say, storage, right? Cloud storage. I'm a user. I start producing a bunch of material. I start consuming more storage. But I'm a human. I can only make so much at same speed.

[00:07:24] Right now, with a single click, you could launch almost an infinite number of tokens to be launched by a machine agent, which just spins off and starts running unconstrained. I look at this and sort of say, like, that variability is going to be really problematic. I think, you know, and even more so in ungoverned, you know, the smaller you are, the less good you are at controls kind of world.

[00:07:51] I think about, like, you know, letting loose one of these agents. And I've been thinking a lot about the chain of how that'll get passed down, right? You know, the LLM providers, the frontier models, they're happy to pass it on. Now it looks like, you know, Microsoft or the hyperscalers are also going to be happy to pass it on. I would expect the vendors using it in their products, happy to pass it on. Somewhere in there, it hits the provider. Who are they going to pass it on or absorb it? And how about the client consume it?

[00:08:21] Like, how are you framing the readiness, particularly when we think about the partner and their customer to do that consumption? Yeah, and I get hit in that story as an analyst because you can imagine I'm staring at spreadsheets with 10,000 lines and I got 10,000 prompts, you know, all the way down the list. And then I've got seven columns of 10,000 prompts.

[00:08:46] I mistakenly refreshed the spreadsheet, which sets loose 70,000 prompts. I just, you know, lost $7 on a mistaken refresh. So this is the token-based economy that we have to think through. And from an economics of partnering perspective, you know, when you take that variable expense based on usage, do you pass that through directly to the customer?

[00:09:11] Or, you know, is that somehow collected in your cost of goods sold and you manage it like you do today with your tech stack and then manage your profitability on top of that? You know, how do we introduce that variability to the customer? We know how it works in enterprise. We know that there's now $1.4 trillion of enterprise credits loaded up at AWS, Microsoft, and Google's marketplace.

[00:09:36] And we know that big customers are making these investments and drawing them down. And there's a huge focus at the high end on FinOps. What used to just be the cost of cloud and the cost of edge to cloud and things is now becoming a token-based economy of how we look at our entire infrastructure,

[00:09:56] the hardware, the software, the services, the cybersecurity, the telco, and manage this across and these new variable expenses that we need to, you know, actively look at, govern, and have compliance and regulatory type of guardrails against. So we don't have somebody down in the organization, you know, burning up data center cycles, you know, without an outcome, you know, attached to it.

[00:10:24] So I know you track a lot of vendor channel program design. Like, are there vendors building MSP-appropriate consumption controls into their programs? Things I think of, like, spend caps or showback reports or operator-facing dashboards. Like, should we be leaning on the vendors? Are they building this? Or is this a gap that MSPs are going to have to fill on their own? It's a gap. So, you know, in the next few years, we have this. And it can be a very expensive one, as we just talked about a few examples.

[00:10:53] Again, a rogue, you know, person. And again, a well-meaning person, not realizing. And, you know, again, 94% of us now use ChatGPT or Gemini or Claude Cowork or something, you know, in our daily lives. Those companies, though, have kind of said, you know, for $20 a month, they're going to just grab the average of consumption. And the people who don't use it, you know, consume nothing.

[00:11:20] The people that are power users kind of, you know, the social part of it, you know, pays for their usage. But it's kind of confusing the market to say, well, I think AI is $20 a month. That's the consumer version. But when we're actually using these advanced models, and again, when we're adding these layers of prompting going to that next layer,

[00:11:44] you know, we're at a point now that we can charge for it, but none of the vendors are at the point where they can operationalize that or build the economics around that inside of a program. So they've left it to MSPs right now to go figure that out in a pretty much a locked price model, $113 per month per device or per person, to try to figure out variability and maintain your profitability. But we're hearing from MSPs that that's very difficult to do.

[00:12:12] We've heard that a large number of MSPs that are struggling with profitability as we enter into this AI era. And again, I think the community here and the areas that we work as MSPs, you know, in these peer networks and stuff, need to figure out best practices and the governed, you know, ability to go in and make sure that we're saving our customers money. This is what we've always done. Right.

[00:12:37] And, you know, we obviously control the risk on ourselves as well in terms of, you know, being able to handle these extra costs and not have a mechanism, you know, to pass them through to the customer. I want to get back to the operator point in a moment, but I actually want to spend one more moment dwelling on the vendor here because I think the other piece that's important to recognize is we're already having conversations that the model providers themselves are subsidizing token usage. That even at their current rates, you know, you said we're figuring out the economics.

[00:13:07] I don't think they've even figured out the economics to the level that we should all feel comfortable with it. When we talk about like Anthropic and OpenAI, we both know that they are subsidizing some of that level. And in fact, their most expensive customers are costing them even more. Isn't this like causing a bit of a wrong motivations for people to consume more tokens as they will start flowing through the system? I mean, the methodology I think about for the top end where they're kind of weirdly subsidized

[00:13:36] and then there's going to be some strange dynamics that come to the vendors themselves who have a potential markup capability as they deliver it through. Like, isn't there something broken over on the vendor side of the consumption of tokens? Yeah, it's a very different market. So, you know, we now know that Anthropic is going into IPO, but we haven't seen any of the official documentation yet. And that'll come in the next upcoming days. We're watching as OpenAI is doing the same thing.

[00:14:04] We watched as NVIDIA became the most valuable company in the world in this next era. But understanding this $7 trillion build-out that's happening now, which doubles the size of our industry, you know, in the next three to five years, all of that is being placed now. Investors are making bets on where those economics are going to flow. So you're going to see Anthropic IPO at a multi-trillion dollar level

[00:14:29] become one of the biggest companies on the planet in only, you know, months of build-out. Now, they hit $45 billion in revenue faster than any other company in history. And it was measured in days in terms of getting there. So this is all unprecedented. And again, the big trillion-dollar questions and the trillionaires that are being created at the top, now there's this big filter of where does the $7 trillion go?

[00:14:56] How much of that does go into Datacenter and NVIDIA? You saw earnings this week, you know, for example, Dell and HPE and Lenovo, you know, having ridiculous-sized earnings. You have the distributors in double-digit growth, and they're making, you know, 60%, 70% increases in profit. So we're just watching now. It's a breaking story in terms of where the benefits lie, the anticipated benefits, you know, as we go here.

[00:15:24] And what we don't want to be as MSPs is kind of left to the end. Duck in contracts that limit our profitability, limit our ability, which isn't being seen in hardware, software, and consumption. But if we're limited in terms of our business model and our ability to, you know, play in this large, very profitable change in the industry

[00:15:52] this next 20 years, this is concerning. So again, I think this is front-page news, and I think that we're watching almost every day, every press release, every earnings release, and the guidance that we're getting from vendors in terms of how their partners are going to participate and how all boats will rise. And that's where I would be critical to say that we're not hearing, you know, specifics. We're hearing from them how they're going to do well in front of their investors,

[00:16:22] but everyone else, the consultants, the designers, the implementation, the integration, managed services, the hundreds of services providers, services, and obviously the hundreds of thousands of service providers is where, you know, kind of the buck stops. Yeah. Now I've got two questions from the community that I want to ask, because one in particular, one of the members of the SBT community made a really sharp point to me. And they said that the vendors that he's working with

[00:16:51] are actively encouraging AI adoption, specifically citing IT Glue has a new AI assistant that pops up constantly saying, use me, use me, use me. Like it's continually like aggressively. And they're training clients, both on the MSP side and in users, to run up usage before the MSP necessarily has any control infrastructure in place. What I want to ask is, like, is that a pattern you're seeing across the vendor base? Is it like a majority? Are they all driving this way?

[00:17:20] Or is it, or are you thinking it's more isolated? Like, give me a sense of how widespread this is. Yeah, there's kind of two answers to that question. I mean, the first answer is there's 250,000 ISVs, like IT Glue, chasing AI. Right. And they're almost announcing vaporware before, you know, it's released in their product. But this idea of this SaaSpocalypse that's coming, that's going to attack a lot of the software that's out there today. Again, 250,000 companies are at risk

[00:17:49] in terms of AI kind of replacing what they do. They're reliant on a subscription. They're reliant on humans to do things. What happens when agents start doing things? The side door access, where you don't need a UI or UX, you don't need a user ID and password to get into IT Glue, but there's data in there that's very, very valuable. And there's workflows and processes and business logic in there that's very valuable to other agents that need to pick that up.

[00:18:19] And it's valuable enough that they're going to pay a nickel every time, you know, they knock on that side door. So in a SaaSpocalypse, every company that doesn't have this data that's very useful, in other words, if it doesn't change all that often, AI can come and read it once and never have to come back logic or the processes are weak, AI is going to be able to reverse engineer it, you know, within milliseconds and figure out that I don't have to pay you a nickel anymore because your data isn't changing and your workflows are pretty, you know, linear.

[00:18:49] So in that case, the SaaSpocalypse says that I never have to come back and a human never has to log in. So everybody's chasing this AI control plane to say that I want to be friendly in the future and I want people to log in so I can charge them a subscription and I also want to charge micro-consumption to these side door access for every other tool, you know, to pick up this critical information along the way for it to be able to drive its outcomes.

[00:19:18] If you're not in either of those camps and whether you're IT Glue or you're ConnectWise or Kasei or anyone up and down the stack, this is a frightful moment of where you're going to be three to five years from now. You're 100% right. In fact, I'm having specific conversations with members of the community who are telling me like, oh, I'm just replacing this smaller vendor, you know, particularly the ones that are where I could make an argument of that's a feature more than a product. You know,

[00:19:48] like this is a small, you know, small offering. It does one particular thing. Well, it's a small tool. MSPs are telling me, oh, I'm just replacing that. Like I can just go have AI agents go ahead and manage that for me and get rid of that monthly recurring fee, particularly if the vendor is aggressively priced. So I'm already hearing that directly from the field. The other one I wanted to get a sense of is when I had this conversation, again, with some of the members of the SBT community, one of the things they push back on is how fast

[00:20:18] it'll ripple through. So for example, I cite the E7 licensing as something that it's the first mover here. And what they've pushed back is said like, hey, you know, E7, that's an enterprise product, right? That's happening. There's at the enterprise, it takes years to reach the SMB and the typical, you know, rollout in a smaller level of customers. Are you expecting similar pacing? Is it going to move faster? Like give me a sense of your thinking about the way

[00:20:47] the timeframe will go, particularly as we're trying to look at these historical analogies. Yeah. The interesting thing that is, coming out of all the research now is one of the surprising things in the AI era, it has so much to do with segment, your customer size in terms of employees. Not so much industry, not so much geography other than some sovereign, you know, type of conversations, you know, not so much on the product stack and the service stack,

[00:21:16] but the segment drives everything. I think one of my most popular posts this year on LinkedIn was breaking down the segments. The, you know, MSP community that we serve, you know, really have a target audience of about 50 employee customers up to about 150. And in that, very much that segment there, there's almost a 0% chance that any of those customers will be calling Sam Altman or Dario at Anthropic or NVIDIA, Jensen

[00:21:45] to go buy a GPU. they're just not in that market. So when you take a look at AI, you know, from an enterprise or a large public sector, you look at it in mid-market and then you look at it in SMB, it looks remarkably different. And this gets to your conversation around tokens and microconsumption and how it's going to play out. So let me just bear with me for a second. In the target market that we have between 50 and 150 employees, there is

[00:22:14] 1,012 1,012 categories of company. You can run a flower shop, you can run a daycare center, you can run a tanning salon, or by the way, you can run an MSP. Out of the 338,000 MSPs, 98% of them sit in that target. So when you think about running one of those categories, you run a flower shop, now there's seven or eight horizontal applications that have been built in the flower shop era.

[00:22:44] They got built on Microsoft or they got built on AWS, but that's what you buy horizontally in your company and that's what runs your sales and marketing, that runs your operations, runs your finance, runs your HR, even runs your product. This is the same rinse and repeat for tanning salons. You know, there's eight major horizontal applications you can choose from. Again, soup to nuts. At that level, you don't buy Salesforce and ServiceNow and Workday and Marketo and NetSuite,

[00:23:14] HubSpot. I mean, you're not buying these big SaaS applications and paying a hundred and something dollars subscription for each. You end up buying a horizontal piece of technology. MSPs, ConnectWise, Kaseya, Enable, Ninja, Halo, Synchro, SuperOps. I mean, all the way down to this, 25 horizontal applications run pretty much our entire business. In that world, agentic AI is going to come to you through those applications in most cases.

[00:23:44] There's a few MSPs that are, you know, working on open claw during the weekend and they've got all kinds of really interesting things going on. But for the most cases, and our customers, they're in wait mode. And the way it's going to be presented to them through these applications, i.e., you know, you're running a restaurant, you have Toast or Square or Clover, which is all the iPads people are running around with. When they see a spike in Uber Eats orders, it's automatically

[00:24:13] agentically going to increase your food order and staff your kitchen. So a restaurant owner isn't going to have to spend 24-7 trying to find these signals. Agentic AI is going to come through Toast. Now Toast is going to build through AWS their backbone using Anthropic or OpenAI and obviously Bernstein up NVIDIA chips, but that's never going to be in front of the restaurant. So when I'm an MSP serving that restaurant, I'm starting to think that the

[00:24:43] actual tokenization, every time that Uber Eats thing gets found and all the agents go to work, the restaurant's not going to be charged a dollar for every of those things. It's going to be consumed inside of Toast, which already has a subscription, and then Toast may charge extra usage fees if you ask it to do more things. The MSP that's managing that, perhaps building additional agents and building all of the security and the data and the

[00:25:12] layers above it, compliance layers above it, is where we're making the money. And then the extra agents that we're writing may be consuming those extra tokens. We have to figure out from a product perspective, not a services perspective, how we're going to pass on that price to the customer. Okay. Now, I've been working on a thesis that I think is worth extrapolating out here. I'm going to do that right after this message from our sponsor. The SMB online conference

[00:25:41] is June 23rd through 25th and registration is open now. Three days of practitioner-focused sessions. Pricing, M&A, AI, private equity, service delivery. No vendor speakers, no fluff. The theme is profitable is enough. If that resonates, you should be there. Small Biz Thoughts community members get in free. Everyone else, $399 at smbonlineconference.com.

[00:26:11] All right, Jay. So here's the thesis that I've been building. MSPs have built their entire business model on the idea that their cost structures are knowable, right? Flat free delivery works when you can model the inputs. The moment a key vendor moves from per seat to seat plus consumption, there's a cost component that's theoretically zero to infinite with machine speed agents as the consumer and deals with it and there's no vendor side motivation to cap it.

[00:26:41] And then the second layer is the governance liability. As AI agents live in all of these systems, that's provisioning, remediating, changing configurations, someone is accountable for what they do. The vendor sells the capability and limits their liability, putting the MSP closest to the environment. So when something goes wrong and the client asks who approved that, the MSP becomes the answer. And I think that accountability is not currently priced

[00:27:10] into any contract that I've seen. So my thesis really and the conclusion I'm coming to is the MSPs are building that control layer now. Consumption variability, spend caps, approval gates, audit trails. They're going to be the ones that are going to be able to sell it as a product. And those that don't are going to absorb it as unpaid liability. And I think the window to get ahead of that is closing. So from your vantage point, you're looking at the channel, like, is that assessment right or am I completely

[00:27:39] overcorrecting here? No, the thesis is absolutely correct. The great thing about managed services, different than other services, I mean, when you sell consulting services or implementation services, you're kind of charging for the service. MSPs have always charged for the outcome. You know, in terms of managing environment, controlling costs, and obviously controlling the chaos that comes in any edge to cloud kind of IT environment, MSPs have really sold that outcome. And in AI,

[00:28:08] it's not a far leap to sell the outcome, obviously, with different infrastructure underneath it. So this idea of, you know, micro consumption or these token based, one of the changes, though, will be a move towards marketplaces. You know, in a world of selling hardware and then collecting the money, you can fax in that order. In a world of subscriptions, you can have direct access to the customer and have some sort of system to accept that monthly fee and to make, you know, moves,

[00:28:38] adds changes to those contracts and things. But when we're charging a dollar per thing, whether on the technology side or the services side, and whether that's a human driving that outcome or doing that specific tactic or an agent, it really doesn't matter. The outcome matters. So we have to maintain our focus on selling on outcome as opposed to all the steps that went into it or if humans were even involved. And we have to, again, keep focusing

[00:29:06] on the outcome-based pricing. But we're going to be surprised that you can't manage that by a fax machine. You can't manage that on some direct ordering system that you send out an invoice. You're going to have to have marketplace functionality and that's going to come either via the hyperscalers. It's going to come and there are all their recent programs to do that that MSPs can take advantage of. Or it's going to come out of the major distributors. You know, think of the TD Cynics and Ingram and D&H and others, Arrow and others.

[00:29:36] It's going to come out of a major platform, a multi-billion dollar platform that allows for the metering of this. You know, when you pay your water bill, when you pay your electricity and things, it goes to a major utility that can meter it on the side of your house and charge you appropriately. The same things is beyond the capabilities of the average MSP to create a metering system. So we're going to have to rely on a marketplace where the customer might actually be buying other things.

[00:30:05] That's where they might buy Toast or that flower shop software. That's where they might be consuming other stuff. And for the first time, we have a storefront that actually is backed by a major distributor, a major marketplace, or a company like an AppDirect or a marketplace development platform like Pax8 or SureWeb that can handle this new economics for us and make sure we maintain our profitability and make sure we're charging the customer for all of the outcomes that they get today and the new,

[00:30:35] you know, better outcomes that they get tomorrow. But aren't the incentives a little wrong there in the marketplaces? So I love your leaning into utility like water and power, but those are regulated, you know, managed monopolies. They aren't just free market like consumption and isn't the financial motivation of those that are selling the metered solution now, like they're incentivized to sell as many of those as possible. Talk to me about the financial incentives because I'm just feeling

[00:31:04] like this doesn't line up. Yeah, I mean, if you leave your hose running all month or if you, you know, go and build a Bitcoin, you know, crypto engineering thing in your basement, you're going to see that, you know, while they're regulated, you're still charged on usage. Sure. And it's not too many months in a row are you going to leave your, you know, hose running all month long or your toilet, you know, leaking all month long. So there is obviously

[00:31:33] controls built in that and there's controls that are built in this as well. And it's just basic supply and demand economics that, you know, if you're losing tokens out of a employee over here, you're going to want to go and manage that and ask the FinOps questions and ask your MSP to go help me control the risk to help me control this environment or govern this environment so that that never happens again. Somebody's checking the taps and somebody's, you know, checking that no one in their cubicle has set up a little Bitcoin mining,

[00:32:03] you know, session. So this is what's happening in our industry. This happens to happen quick and we probably, you know, need to start talking to our customers as they dabble in AI, perhaps even at a consumer level that we make sure that they understand, you know, what could come and the risk they have as an organization that, you know, somebody leaves the tap on. Gotcha. Well, I've got to throw this to Harold. He's got a great question. Jay, he's asking, is customer trust so powerful to fend off

[00:32:33] these AI first upstarts that'll come in and just give away software to build market share? Is the customer trust strong enough? It's interesting. That's the way the economy, if I look back on 56 years of our industry, I mean, I just finishing off Amazon's book and, you know, the fact that they could lose money for the first 25 years and, you know, most SaaS companies never made money for the first, you know, 10 or 15 years until investors stepped in. And this is how AI is just reliving

[00:33:03] that playbook. Anthropic and OpenAI, I mean, I guess Anthropic has got a road to profitability, but OpenAI is losing buckets of money and it doesn't matter because everybody visualizes the $7 trillion getting your fair share of that and then at some point we'll figure out how to charge each user, you know, a nickel and like Facebook or one of these other big network effect companies, almost within days you can become, you know, a trillion dollar company

[00:33:32] by just turning on the commercial engine when it makes sense. There's books like the in-shittification of networks that go deep into this in Uber, Airbnb and every consumer example, but it's going to happen here too. Every one of these companies is motivated to lose money, to gain users, gain the network effects and to the question, this customer trust, to earn that customer trust that, boy, is this network a great place to be. Uber is so much better than a taxi

[00:34:01] until it isn't and that's where the, they switch and flip the switch on commercialization. MSPs are the trusted advisor. That's where the customer trust has been built and as part of this is to play out this future of network effects around AI and where the, you know, money could go and could get lost is absolutely a service that MSPs should be selling their customers today which is called FinOps.

[00:34:33] Gotcha. Now, so I guess as I wrap up our time here, like is that where you think MSPs right now, you know, we're talking in June of 2026, is that the portion that they should be spending the most time getting ready? Particularly because some of your advice was, hey, we're going to have to wait for pieces from the larger providers. We want to give them some tactical advice. Is that the best use of their time right now? Yeah, I mean, the advice I think for the last few years that, you know, I've been given based on research, we've had the chance

[00:35:02] on the top 25 tech companies in the world to go, you know, talk to every single one of their customers and every one of their partners and build these multiplier reports. You heard the $7 you can make for every dollar of AWS or the $8.45 you can make for every dollar of Microsoft. You hear it at CrowdStrike now, you hear it at Cisco, you hear it all through the, you know, every major vendor in the world. But to get those multiplier numbers, how much of that is consulting, how much of that is, you know,

[00:35:32] designing or implementing, how much of that is managed services. The great MSPs now, there's not a race to get to 100% recurring like there might have been 10 or 15 years ago. The race is for customer value and customer trust driving those outcomes. And if I could make $2 or $3 for every vendor dollar by showing up before, during, and after the transaction and managing it, some of that's going to be project-based revenue and it's going to show up in different ways. But if I could make, you know, $3 or $4,

[00:36:03] that's the level of customer trust. And I'm not just chasing managed services revenue, but those other revenue items I want to operationalize and become monthly. But until they do, I want to be making all kinds of revenue, adding value at every stage of the customer journey. And that's really the advice is the best MSPs in the world are not 100% recurring revenue. There we go. That is the piece of advice to end on. Jay, I really always appreciate your time. Where can people

[00:36:32] follow your work best if they want to learn more about the research you're working on? Yeah, best place is probably LinkedIn. I make about a daily post and as we generate new research, I share most of it. And if you send me a quick note, there's probably a 30-page report behind it. I can slip over to you and we're looking to raise all boats here and, you know, share it as much as we can. Well, Jay, I know you're a great contributing member to the community. I always appreciate having you on. And for the audience, Jay is bringing his full analysis skills

[00:37:02] to the SMB Online Conference on June 25th in Session 10, Channel Intelligence and Independence. If you want to go deeper into where the channel's headed and what it means for your business, be in that room. Registration is live at smbonlineconference.com and Small Biz Thoughts community members, your free passcode is already in the community. Jay, thanks so much for joining me today. Thank you. And I want to thank our sponsors for making the Business of Tech Lounge possible.

[00:37:32] ABC Solutions, if you're running an MSP and struggling to get clean financial visibility, that's profitability by client or service line performance or just accurate books, they focus specifically on accounting for managed service providers and IT firms. They're at abcsolvesit.com. And Rhythms, if connectivity reliability is still a constraint, especially in edge environments and for distributed teams, they deliver 5G solutions designed for MSP use cases where uptime and coverage

[00:38:01] actually matter. More on them at R-Y-T-H-M-Z.com. If you want to go deeper on any of this, daily analysis, operator level breakdowns or the channel chatter intelligence offerings that we have, head to businessof.tech slash plus for Business of Tech Plus membership. Thanks for joining. I'm Dave Sobel. This is the Business of Tech and I will talk to you next time.

[00:38:47] Produced by Picture This Video, part of the MSP Radio Network. of the MSP Radio Network. Thank you. .