Triangulating Internet Growth

In yesterday's post, we saw that the big Internet backbone providers like Level 3 Communications really want to see a boom in the rate of Internet bandwidth consumption, and would be thrilled if that boom was accompanied by a sudden reduction in the rate of bandwidth price declines.  More units sold for more money per unit equals bigger profits.

So, are we heading for a boom?  To get a read on that, I looked at Level 3's "IP and Data Services" revenue for 2007 to see if I could spot a trend.   Now, looking at a single company, like Level 3, is kind of dangerous when trying to extrapolate a trend for an entire industry, so take this analysis along with a disclaimer that, well, it could be wee bit speculative!

Here was Level 3's IP and Data Services revenue:

	1Q07	2Q07	3Q07	4Q07
Level 3 IP and Data Services Revenue ($millions)	$     144	$     143	$     144	$     153

You can see that revenue was really very flat until the fourth quarter, when it saw a 6% quarter-over-quarter jump.  You can interpret this data in one of two ways:

• Look at the entire year, which gives you an 8% annual revenue growth rate, supporting the hypothesis that price compression is eating most of the revenue you would otherwise achieve from bandwidth growth, or
• Look at just at the fourth quarter, and assume every quarter thereafter will grow just as much, and you get a much prettier 25% annual revenue growth rate.

Okay, assuming 2007 is a good guide for 2008, that gives us a range of 8-25% revenue growth in 2008.  So, how much bandwidth growth does that imply?  To find that, we need an assumption on how much prices will decline in 2008. 

Again, looking at 2007 as a guide, if you assume that Level 3's 2007 bandwidth growth matched the overall Internet's bandwidth growth, then you would have to assume that Level 3 grew 50-60% in bandwidth in 2007.  Some sources think that Level 3 officials have dropped hints that the actual bandwidth growth at Level 3 was more like 90% in 2007.  So, we have a range of 50% to 90% bandwidth growth in 2007.  This much bandwidth growth, combined with what we know about Level 3's Internet revenenue, results in a range of price compression between 28% (if bandwidth grew 50%) and 40% (if bandwidth grew 90%), year over year, for 2007.

What does all this mean for bandwidth growth in 2008?  Well, if 2008 is similar to 2007 in terms of revenue growth (8 to 25%) and price compression (28 to 40%), then you get a range of potential Level 3 bandwidth growth as follows:

Likely Level 3 2008 Internet Bandwidth Growth Range
	Price Compression
Revenue Growth	28%	40%
8%	50%	80%
25%	74%	108%

So, you're probably thinking "Thanks a lot, Ike.  So we know it should be somewhere between 50% and 108% bandwidth growth in 2008.  Thanks for narrowing it down!" 

I know this isn't much of a prediction, but I wanted to take you through the calculation of how you end up with this range of performance, because the analysis might indicate that an acceleration in Internet bandwidth growth is already underway, at least at Level 3, and it may have started in the 4th quarter of 2007, as shown by Level 3's revenue data.  In just one example, if you pick the midpoint of the 50% to 108% range, you get a 79% year-over-year bandwidth increase in 2008, which is certainly a better number than the standard 50-60% we have seen for a few years.  If that accelerated growth rate of roughly 79% sustains for a few years, I would certainly call that a resurgence!

Of course, to use this as a prediction for the whole industry, it helps to believe at least six things:

• That Level 3's revenue growth rate is a good proxy for the whole Internet bandwidth industry,
• That Level 3's price compression rate is a good proxy for the whole Internet bandwidth industry, and
• That Level 3's bandwidth growth rate is a good proxy for the whole Internet bandwidth industry.
• That Level 3's 4th quarter jump in Internet revenue was not just a one-time seasonal blip, and that some of that jump was an accelerated growth rate that the whole industry is experiencing, and that will extend into 2008 and beyond.
• That Level 3's 8% growth rate in Internet revenue in 2007 represents a floor for 2008 Internet revenue growth (after all, it could be lower, couldn't it?).
• That a slowdown in the US economy won't create too much of a drag on growth.

Any of these might be an erroneous assumption, so as always, use this little exercise at your own risk!

PS - click the continuation link to see the math that relates revenue growth to bandwidth growth and price compression.

Postscript: Here is the math of how you figure out annual revenue growth rate.  If you know

• Annual Internet bandwidth unit growth rate, and
• the annual rate of decline in Internet bandwidth prices

...then you can compute

• Internet bandwidth services revenue growth rate

In fact, if you know any two of the above numbers, then you can compute the third number.  Here is the formula:

Internet bandwidth services revenue growth rate = ( (1 + annual unit growth rate) * (1 - annual rate of price decline) ) - 1

Comments

[Darren Loher]

I think you've got the formulas right!

However regarding the speculation on traffic growth: from my experience looking at only one provider is not completely bad, but that one provider must have a very large Internet backbone (probably a top 5 by volume provider, which Level 3 most likely is). The other critical factor is that the measurements must be taken over a long period of time to smooth out bumps due to one-time and/or special events. In this case, I feel strongly that 4 data points in one year is not enough evidence to indicate a trend.

I would not take the Q4-2007 bump as indicating a trend. A special event is more likely the reason for the bump. For example, when a customer deals is signed or a new product is launched, traffic tends to shift from one provider to another. While I don't have any information to back this up for Level3's Q4-2007 revenue, I suspect this is the more likely case than traffic grew that much over the Internet at large.