Does Market Share Matter?

Yes, but only a little.

When it comes to operations it is important never to be too far off the beaten path. Every customer site is unique, so then how do we tell if we are on or off the beaten path? Market share reports provide us this information. Most share reports have to do with vendor sales performance. Some look more to the installed base and show what's actually out there. These reports are valuable, but don't put too much credence into them and don't allow them to dictate a strategy.

First a little history. Telecom used to be all about ports. Every phone, every trunk, every TIE line, and every other connection required a dedicated port on the system. Some ports were built-in, and others came on expansion cards. Hard to believe this was true just a short time ago. Ports had a real cost associated with them (both material and physical space), so organizations tried not to buy unnecessary ports. Nor did vendors give them away. Ports were tangible and easy to count - that is, there wasn't much debate over what a port was.

For years, ports served as an important aspect of market share. The total number of ports sold were one indicator, but total ports divided by total customers gave us an average sized customer. Total ports in inventory told us about manufacturing and supply. Those were the good ol' days. Unfortunately, the ports have gone out to sea.

In an IP software-defined world, we no longer can rely on ports. A single IP connection can support thousands of "endpoints" (phones, endpoints, mobile devices). So what do we count now that the ports are gone? This is where it gets complicated, painful, and humorous.

The good news is there are several alternative things to count instead. The bad news is none of them are as good as ports. Phones, endpoints, extensions, revenue, and licenses are the top contenders - each with their own set of problems. Phones are problematic because of third party and soft phones make it difficult to correlate phone count with system size. Also, phones work on old and new systems, so new phone sales don't confirm net new growth.

Extensions are also tough to count as most vendors don't technically restrict them. There are physical extensions and phantom extensions. Some systems consider each branch of an automated attendant to be an extension. Speed dials or paging codes can be extensions.

It would be nice if we could just compare vendor revenues, but that isn't particularly meaningful any more either as few vendors offer near identical products. What vendors bundle in their prices all vary. Not to mention, not all vendors break-out their revenue details making comparisons extremely difficult.

Licenses seem to offer the best alternative to ports, so that's what mostly gets counted. But licenses are not real physical like ports were. Adding more licenses does not make a system physically bigger. Adding unused licenses does not necessarily cost the vendor anything. So naturally we've seen bundles and packages that equip users with many more licenses than they need. This is why most market share reports still count telephony licenses (do-able) instead of UC licenses (impossible).

I do respect and admire market share reports. But please understand no two come to the same conclusions based largely on the same data. However, because they are generally consistent with themselves, patterns and trends do emerge over time. You can also see a general pecking order, but that's about all I get from them.

The market share reports are generally sold to the vendors, and vendors like to twist them a bit to make them look large. Here are some examples:

Cisco:Cisco would love it if we would just count IP phones sold. They sell a lot of phones. This is partially because many hosted firms use Cisco phones. Also, many Cisco models are available over the Internet at decent dotcom prices. Most competitors restrict their endpoints to authorized dealers only.

Microsoft: Microsoft wants us to count licenses. They make Lync licenses (standard and Enterprise) available through Enterprise software maintenance plans. Many enterprises obtain rights to use Lync for no additional charge. To license Lync for voice requires still an additional license - yet Microsoft rarely separates those figures; instead, they generally refer to customers that have deployed Lync (both with or without voice).

Avaya:Avaya likes it when applications, specifically contact centers, are given considerable emphasis. They also like big studies that compare total revenue and market share as opposed to smaller SMB or enterprise only segments. Few competitors are in so many markets with so much history and installed systems.

ShoreTel:ShoreTel likes it when we count pure IP users. ShoreTel systems never supported digital (and most of its competitors did). So while competitors are selling hybrid systems, ShoreTel is pushing "pure IP." Additionally, ShoreTel is strongest in the U.S., so pure-IP U.S. sales are something they like to share. They share that category primarily with Cisco, so limiting it to say companies with less than 5,000 users puts ShoreTel on top.

Conversely, NEC has done well worldwide, so prefers market share reports that reference global sales. Mitel, Aastra, and Siemens Enterprise Group all have dominance in specific regions (UK, Nordics, and Germany). They prefer market share reports that focus on specific regions. Digium Asterisk is wildcard as their Asterisk product isn't sold. Digium prefers to quote the number of downloads, which is impressive.

The point here is that depending how the data is diced and sliced, outcomes vary dramatically. So it is important to view these market share reports with a critical eye. I argue what is more important is feature-set matched to requirements. For a litmus test look at share growth, or possibly a review of the financials and check for growth or profitability - ideally both. I don't think buying #1 means a whole lot, at least not when compared to buying the best solution.

Dave Michels is an independent analyst at TalkingPointz.com.