Skip to content

IBM SMB Analyst Summit Day 1: "IBM is great! …but it's not for me"

IBM has a problem: their brand and technology are strong up and down the market, but it’s common knowledge that only large enterprises buy from IBM. That tale was told many ways during each session this afternoon by IBM execs and even customers and partners.

Sure, they’ve been working on it for several years now: making the mid-market and (to a lesser extent it seems) the small-market think of IBM when they need technology and business services. But, everyone still realizes that there’s still work to do before the IBM brand has as much prominence in SMB mind-share and purchasing decisions as, say, Dell.

The strategy so far has been three pronged, each revolving around simplifying IBM’s technology and services:

  1. Par down enterprise offerings into the Express line.
  2. Grow the partner ecosystem to get more partners selling to SMBs.
  3. Re-work internal IBM processes like financing to accommodate SMBs.

I’ve always been a bit flummoxed of IBM’s partnering strategy in the SMB area. My concern isn’t that they do it poorly — according to what I’ve heard today, they’re quite good at it. Rather, that they seek to do so much of their SMB selling through partners instead of directly themselves. Their core competency in this space then becomes creating the technology and services that partners sell and then maintaining those partner relationships.

What’s in their blood?

The question in my mind however is what IBM’s giving up in tacit and operational knowledge by out-sourcing so much of the relationship with the end-user and customer. The Agilist in me values disintermediating the relationship between technology producers and technology users; it seems like IBM is trying to mediate that relationship as much as possible.

Now, the benefit on IBM’s side is one of margin boosting, risk management, and acknowledging their core competency:

  • Using partners as a sales channel, IBM “only” has to spend money on successful deals. They don’t need to spend cash to beat the street finding all the frogs before they find a prince. More importantly, IBM’s built up a bevy of technology and business processes to sell over and over again with a minimal amount of customization.
  • As the SMB market swings — or gets “spikey” as Steve Solazzo put it — their partners can absorb part of the shock instead of IBM taking all of it. Don’t get me wrong, IBM would suffer, but at least it’s spread out.
  • With the sell of ThinkPads to Lenovo, IBM’s gotten rid of all it’s consumer-oriented areas. As Steve reminded me last week, the last mass-market play IBM did, OS/2, didn’t go so well. Corporate scars run deep and rarely are forgotten.

As with most everything IBM does, the lay-up to the conclusion is well reasoned. That is, if you accept the desired constraints, the plan looks sound.

The problem, then, is in those constraints. In the case of IBM, the two problematic constraints are:

  • We don’t sell “applications.”
  • We don’t want to be as far away from consumer oriented sales as possible.

In my mind, the SMB market is all about the opposite of those two things:

  • SMBs don’t care about middleware and hardware, they want applications.
  • SMBs want software with the minum amount of fuss and relationship with a vendor or provider.

And, to Jon’s comment from this morning, they want it cheap.

Also, part of the problem is that SMB to IBM spans a large amount of customers. That said, if their diction today is any indication, they’re more interested in mid-market than “small market.”

But, let’s get the point about applications. In the area of application sales in the SMB market, IBM has competition from (at least) two camps:

  • Open source companies and startups.
  • Hosted, or SaaS, companies.

In both cases, the companies are selling applications rather than “raw software.”

Open Source

Open source companies may, like IBM, have as their core asset middleware, but what they sell is helping making that “raw software” into an application. They don’t sell the middleware itself as IBM does with licenses software. The distinction is subtle, but it gets down to price: customers are paying for applications rather than the middleware that those applications run on.

Now, IBM has quite a foot in the open source world…but I’m hesitant to think that IBM sales would jump at the chance to give it away middle-ware (choose open source) rather than license closed source middle-ware.

Open source companies, of course, don’t mind taking less. To them, the only direction to go is up, so they don’t have to worry about cutting out middle-ware licensing fees. They realize, and are realizing, that software itself has less and less value as open source middle-ware matures.

As I come back to frequently, the question becomes: can that theory of software value support the size of companies like IBM, Microsoft, Adobe, and other companies that were born in the closed source, licensed software world?

While IBM has been adopting to monetizing the “configuration” and support for software, it’s still odd to me that they’re not moving more aggressively to in-house more of that. It seems like the only hedge against the idea that selling software is an increasingly difficult revenue source.


The other competition to IBM’s SMB strategy are what you might call “pure application” models seen most prominently in hosted/SaaS companies like SalesForce. Here, from the customer’s perspective, the middle-ware is the web browser. That is, there really is no middle-ware.

Since the beginning of time, the industry has tried to move to a hosted model, often with abysmal failure. Each time, it was a question of technology being too slow and unreliable. Before the current go at hosting, ASP models where thick and clunky GUIs were hosted over the Internet fell flat on it’s face. The desktop and local networks were always faster.

At the moment, it seems like the hosting-hopefulls finally understood the core issue: if you’re going to run a hosted application, it has to do less and be less flashy. Web-based applications, with their limited methods of reading and writing data and pushing along work-flows unintentionally provided the technology for this constrained view.

As an side, much of the RIA technology that I see seems to gloss over the important, and completely accidently, lesson of the web: once the network is involved, you can’t do as much. But, without the network, your applications is boring.

So, why is this a problem for IBM’s SMB efforts? In short, the answer is because a hosted model is the ultimate disintermediation between technology supplier and technology user. From what I understand of IBM’s SMB model as outlined above, dealing directly with end-users and customers isn’t something IBM is interested in.

Thus, if/when customers demand more and more hosted offerings, IBM will have a skills and experience gap to fill.

Now, that would have been my position without exception until we talked with Hoplon in a RedMonk Radio podcast a few weeks ago. What was remarkable in there was that IBM was selling mainframes with a Linux/Java facade on-top of them as pure hosting. While that isn’t selling hosted applications, what it highlights is a skill that many people may have forgotten that IBM has: running your IT. Granted, that skill seems to be focused in mainframes, but it seems like IBM could transform it into selling a beefed up, IBM backed version of Amazon EC2.

IBM could offer applications on that grid, and, in a way that subverts my point about partnering, IBM could offer that space to partners for selling applications.

Still, again in this case the gating question is: can IBM make enough money in that model?

Sidebar: Brand

There was much discussion of brand during the Q&A’s, revolving around how IBM’s strong enterprise brand hampered their SMB sales. As the title of this post implies (a quote, more or less, from IBM’s SMB research), SMB people know and “trust” IBM, but they don’t think IBM’s offerings apply to them.

The problem is that IBM is having a hard-time with brand extension, using the power of the brand to extend into new markets.

Two “fixes” for this problem could be:

  • Completely re-brand the SMB offerings instead of just adding the “Express” suffix to enterprise brands. For example, IBM WebSphere Express would drop the WebSphere name and take on a completely different “base brand.” Now, that may seem like crazy think, but other companies — particularly car companies — seem to do OK with it.
  • Focus on coming up with completely new features, the killer apps, for SMBs. Then the brand will follow. My conception of Express offerings is that they’re just simplified versions of their enterprise siblings. Instead of simplifying something, build it from scratch to fit the SMB need directly, from the start. I know that “simplified enterprise offerings” understanding is not entirely true esp. when if you pull in the Community Edition/Open Source equivalents. That said, I don’t think the CE brand/line was mentioned once today.

I’d put my money on the second, but it means a large R&D and/or acquisition spend on IBM’s part. Along those lines, the first thing I wrote in my notes when thinking of this brand extension problem was that IBM needs to partner with or acquire a pure consumer company: someone who only knows consumer and SMB sales.

Time for Dinner

Well, I’m a bit late for dinner already, so that’s enough for now. The challenge when I talk to all large, mature software companies always comes back to the same thing: once you discount the price of licensed software due to already and rapidly mature open source offerings, and, add in the pressures from “good enough,” hosted applications, how do you make up that money without hurting The Numbers?

I don’t want to portray that I’m painting a gloom-and-doom, “Death of X” scenario here. Far from it, I actually have a lot of faith in IBM to figure out how to stay alive, indeed, even thrive. They have a good track record. (Speaking of brand: there’s the effects of a strong brand. ;>)

Instead, I’m suggesting that the SMB problem boils down to both coming up with new technologies to match SMB’s wants and new packaging/deployment (like hosted) for those offerings. The margins there are small, and the market model needs to rely on a high volume, which means less customization and lighter, if any relationships, with customers.

One thing to keep in mind is an often referenced fact, including today: the majority of IBM’s revenue is still from a small group of larger, very large, customers. The sort of people who spend $20M each year with them. I call that out because all of this thinking revolved around the opposite of that type of business culture.

Well…we’ll see if there’s a plateful of answers downstairs.

Disclaimer: IBM is a client.

Tags: , , , , ,

Categories: Conferences, Enterprise Software, Marketing, Open Source.

Comment Feed

4 Responses

  1. I wonder if people in the SMB/E space worry about an invasion of the killer service teams. Do IGS have non-enterprise servcies teams? If not, maybe they could buy/partner with the likes of Thoughtworks.

  2. "can that theory of software value support the size of companies like IBM, Microsoft, Adobe, and other companies that were born in the closed source, licensed software world?"

    I suspect the answer to that is largely no, although it doesn't mean that big software companies are going to die out, just that they don't have a good way to get into the SMB market. I have been trying to think of a case of a software company successfully going down-market. Can you think of one?

    Pushing their software through partners is reasonable even from an Agilist/decentralist approach–yes it involves mediation, but it's also a kind of decentralizing action, pushing power and capability and profit out to smaller and more dispersed nodes. Your suggestion of rebranding IBM stuff entirely is a way of pushing power and meaning outwards and away from the center, are there other ways to do that? What about letting the partner ecosystem add new features? Not a complete open sourcing approach, but opening up the source to partners.

  3. wow, lots of good analysis here. Although I am an old IBM guy, even I don't remember when IBM sold coffee grinders and butcher's scales.
    But I do remember IBM's being very strong in SMB – the image of IBM's being only large enterprises is only about 10-15 years old in a 100 year old company. So I think we can
    fix that with time and marketing.
    Interesting also that you commented about hosted applications and environments as an option for IBM SMB – you refer to Hoplon which Redmonk now knows well and it's a very small company – only 40 employees and hosting works well for them

    Tom CurranNovember 7, 2006 @ 5:07 pm
  4. Bill: the answer is a qualified yes: IBM has non-enterprise service teams. For the most part, however, their mid-market or done through partners.
    Anne: successfully going down market…though it's by acquisition, how about Cisco and Linksys? You're question is right though, it's a tough example to fine. Of course, IBM in the PC market until the clones took over.
    Tom: I'm glad you liked it, and thanks for the comment. It was nice meeting in person. And, I think you're right. As many discusions at the event and during 1:1 went, there's plenty in the IBM catalog, it's just figuring out how to package it up correctly and get that packaged item in front of the right people. I look forward to the IBM coffee grider 2.0 😉