Friday, September 4, 2009

Internap CEO details expansion, turnaround strategy

The second part of Eric Cooney's interview with Telephonyonline is available at this link.

Some random comments:

>>On how to spend $50 million: Over the next 18 months, we’ll acquire, build or buy Internap-controlled space in two or three markets. There will probably be at least one expansion of an existing Internap market and probably one new market.

not giving out any detail on the most important thing, location. Let's assume they are still working on a few options. Existing "own" locations: Seattle, Atlanta (but using partner QTS site...), Houston, New York, Boston. Just opened expansions in a couple of them.

>>On the current market for data center assets: Some peer-group data-center providers need to offload assets to reduce their leverage, so it’s essentially a balance-sheet restructuring exercise [for them] that has nothing to do with the market, and they probably would prefer not to be selling those assets.

sounds like a reference to Navisite or other Companies in a similar situation. It remains to be seen if any customer (i.e. revenue) would be coming included with such acquisition.

>>On the impact of the expansion: It clearly will move our mix [of data centers] to [become] predominantly company-controlled assets. It’s about 50-50 [today]. Just throwing rough numbers at you, $20 million dollars [or more] in incremental revenue on an annualized basis could likely come from a $50-million buildout. Twenty million dollars in incremental revenue on a data center business that today is generating roughly $120 million in annual revenue and a 50-50 split [between company-controlled and resold data centers], we go to 80-60, rough-split. That’s the impact, orders of magnitude, of this incremental investment. The more important point is, from a strategic direction, the future dollars of data center revenue our sales force sells will be predominantly company-controlled data centers, which is a significant divergence from our historic approach. The margins we derive from our company-controlled assets are in the same ballpark as [those in] our peer group -- in the 50s. The market leader, Equinix, delivers [more than] 60%. Reselling somebody else’s data center space, the margins are relatively thin. You can see [that] in the gross margins of our data center business – 25% in Q2.

I assume (English not my mother language...) that Mr. Cooney is saying that he expects this expansion to add about $20 million in revenues per year, in the "own data center" category, bringing revenues to 140 million (from $120 million, right now - so we get the $80 million "own", $ 60 million partner sites numbers). That's assuming full capacity for the new expansions, of course. Not clear if this vision includes "transferring" customers from partner sites into the own facilities. Doesn't sound so, if we understand correctly. If he hopes to get from 60% to 80% of revenues in the "owned" data centers category, that's a different story, and includes transferring customers/reducing the number of partner sites. Assuming stable revenues ($120 million), 60% is $ 72 million (from $ 60 million) - 80% is $96 million, which is $ 36 million more than today, and probably needs additional capacity made available to be achievable.

Cooney says Internap "own" data centers enjoy margins in the 50% range (slightly less than Equinix, the best performing Company in the sector, but similar to other peers). If that's correct, as Internap "is generating roughly $120 million in annual revenue and a 50-50 split [between company-controlled and resold data centers]", and "You can see [that] in the gross margins of our data center business – 25% in Q2.", it looks like partner sites are re-sold at cost, bringing a 0 (zero) margin to the mix. That should also include the wholesale deals like 365 Main and QTS. We just can't believe it, or they'd better dump this side of the business.

Going on with this "logic", adding $20 million in revenues at 50% margin would push margins (assuming $140 million per year for colo, and the 80-60 split) in the 28.6% range, assuming the rest of the mix is unchanged.

Still low.

Back in time:

>>Jim DeBlasio

Now as I mentioned on the call, we have a considerable effort to move into a position where we are selling predominantly our own space going forward.

Internap 3Q 2008 conference call, transcript by Seeking Alpha.

Since that c.c., Internap has mainly sold space in partner sites (at no margin?). We'll see if they execute better this time.


Anonymous said...

it is bit confusing but near zero income ( margins cancel by expenses ) does not seem to make sense . some thing in numbers or assumptions here . this is worth an inquiry . thanks . badri

Paolo said...

If you feel like contacting I/R, and share the comments, you're very welcome.
I gave up contacting the Company as I/R was basically useless.
My two cents: 50% margins in Internap's data center is a target, not a reality, today, which, if properly explained, could actually be good news (improving on margins of controlled space and increasing it).
But it's just guessing game...