One of these things doesn't look like the other

I've written before how much I dislike projections. Mostly because they're always wrong. And especially in this economy, I'm dubious of any revenue projections. No one (well, mostly no one) saw this huge downturn coming, no one really knows how long it will last, and no one really knows which sectors will be hit for how long. So there's that. But here's another reason I dislike projections: They very rarely take into account anyone else's predictions. Let's look at a couple of predictions that have been put out there this week: 1. IMS says spend on video analytics might reach $140 million by 2012. 2. Steve Hunt says spend on PSIM software will reach $3 billion to $4 billion annually by 2012. How does that old Sesame Street song go? One of these things doesn't look like the other? Let's take the first. IMS says $50 million was spent on analytics in 2008 (which seems reasonable to me, though I'm always skeptical of such a round number). And they think that might get to $140 million by 2012. So, that's a $90 million increase in four years. It's a CAGR of 30 percent (sick of doing math? Use this handy online CAGR calculator). Now that's a pretty healthy growth rate. But the technology is still young, and the numbers are somewhat small, so percentages sometimes look big in those kinds of environments. For example, here's an article from the government about the geospatial industry growing at 35 percent annually, and 100 percent in the commercial subsector. Here's a nanotechnology company growing at 150 percent plus. Mostly these sorts of predictions are wrong. I don't think anyone believes the access and surveillance market grew 37 percent annually over the last three years. But they're sometimes right. I'd certainly buy that smart phones grew 44 percent year over year from 2006 to 2008. So, let's get to #2. Now, just the language of "$3 billion to $4 billion" gives me the willies. But let's start with this:
A top-down estimation of the potential PSIM spending would normally take total security spending and parse it out based on certain assumptions. For example, of the, say, $50bn annual spend on security products and installation services only a portion would be spent in organizations considered candidates for PSIM solutions, like corporate facilities, airports and the like. Further percentages would be peeled away assuming the adoption lifecycle of PSIM-like technologies in those environments (long sales cycles, integrator reluctance, etc). That sort of estimation could very reasonably find $1bn to $5bn or more of expected spending on PSIM products over the next few years. One magazine published a quote by a vendor claiming $10bn by 2012.
So, it's very reasonable that (using a fairly ambitious 6 percent overall growth rate for the industry during a brutal global recession) of the $60 billion spent on security products and installation that about nine percent of that would be PSIM software? What? Even take the $1b number. Is PSIM software going to be more than one percent of the overall industry spend? That doesn't seem right to me. But, moving on:
How many of those $1bn+ organizations will spend $100,000 or even a million dollars by 2012? To determine that, we interviewed end user executives in 15 more companies and described scenarios where PSIM solutions could produce value (see Table 1). For example, we described specific uses of technology to handle situations involving three or more data types in corporate, transit and government environments. ... For every Global 3000 candidate able to spend $100,000 and more, there will likely be three or four candidates in the much more populous mid-sized corporate or government categories spending $50,000 or more, effectively doubling the total spend of the large organizations.
First, there's a big difference between a company willing to spend $100,000 and $1 million, but say every single one of the Global 3000 spends $100,000 on PSIM software (that seems pretty ambitious, right?). That's only $300 million. So then let's say there are 12,000 of those smaller companies willing to spend $50,000 each (again, pretty good market penetration, I'd say). That gets us another $600 million. So, if the PSIM companies out there today - Intergraph, Orsus, Proximex, Lenel (do they even count?), VidSys, CNL, etc. - deploy 15,000 systems in the next four years we're talking about total spend of $900 million. Double the costs of the systems across the board. 15,000 systems for $1.8 billion. Unless we're including the cost of all the other technology involved in a system run by PSIM software, or we're including all the installation and service, I just don't see how you get to something like $3 billion. And nowhere in Hunt's post is what he thinks those companies' revenues are right now. Doesn't that matter? You'll see the question by John Honovich on the bottom of Hunt's post. He posits their sales at $200 million in 2008. To get to $3 billion, that would require a CAGR of 97 percent. Even if it's $500 million, we're talking a growth rate of 57 percent. Why would it be so much faster than analytic growth? Why would we be spending $140 million on analytics and $3 billion on PSIM software. Aren't there a lot of applications where a couple analytics might make sense but a full-blown PSIM installation might not? The real point of all this is that I just don't see how these two guys could come to these two conclusions independently of each other. They seem wildly disparate. I guess that's why you shouldn't bother asking me for a prediction. What's the market going to be in 2012? I have no idea, really.


One of these things is not like the other. One of these things just doesn't belong. Can you guess which thing is not like the others before I finish this song?

[...] part’s kind of important. Despite some people estimating the PSIM market at anywhere from $3 billion to $10 billion by 2012 (okay, okay, both of those numbers technically originate from Steve Hunt - [...]