Ga. It frustrates me when people do this.
“Check out this chart of the traffic of fotolog.com. They’re now 33 in the world. What’s neat is that the progression from one place to another was pretty linear. No miracles, no interventions, no tipping point or inflections.”
Now, that’s just flat out false as soon as you pull back from the picture a little bit. The picture above shows a six-month window.
Here’s their graph over the last year.
And the last two years.
There actually is an inflection point. A significant one.
So…does anyone know what Fotolog did in on March 1st of this year that fueled the rocket ride? It looks like the site took a hit for a couple of weeks, then came back with a vengeance. Site redesign? Easier-to-use tools? New awareness campaign?
Now the funny thing is, I totally agree with Seth in principle on this point he makes:
“The mistake bloggers often make (actually, all marketers make sooner or later) is the believe that being popular is its own reward. That once every one does their line dance or visits their restaurant or wears their fashion or reads their blog, then it will be popular for being popular.”
A great customer experience, combined with a product for which those customers have a need, will fuel the sustainable, steady, solid growth. It’s unfortunate the example that was chosen doesn’t support that concept with facts.
UPDATE: Per my response to Seth’s comment below, I added the red circles in the charts to highlight the point where something appears to have changed in Fotolog’s business, moving the trend from “flat” (which it had been from 2002 – February 2006) to “growing” (which it’s shown from March 2006-November 2006).
No foul!
I didn’t smooth the chart… that was the default. The thing is, weekends on the web are a mess, and you need to remove them or every chart will look like that.
As for what happened long ago, I don’t know. You’re correct in that it took a while for Scott to get it “right”. My point was that the slope didn’t magically improve… it went from one straight line (broken) to another (not broken)… weekends excepted.
Hi, Seth.
Wasn’t referring to the smoothing; rather, was referring to the mega-hockey-stick-discontinuity-goodness that happened in March of this year. There was something that changed the slope, significantly, in March ’06, and changed the “m” in the y = mx+b slope. Was wondering what it was.
The “weekend effect” is something that we see a lot in the alexa numbers, especially in biz-oriented sites like salesforce.com that have a very predictable drop every Saturday and Sunday, like clockwork. Duly noted.
>As for what happened long ago, I don’t know. You’re correct in that it took a while for Scott to get it “right”.
Bingo. That’s the slow-and-steady message that I think we do agree on.
Have a good weekend.
“Check out this chart of the traffic of fotolog.com. They’re now 33 in the world.”
That would be Alexa’s world, not “the” world. That to me was the bigger gaffe, such as it was.
I agree with you here. I was just going to post the same thing when I read your post. There was a clear tipping point from where Fotolog went from slow growth to huge growth.
Leaving aside the fact that Alexa is next to useless, it’s worth noting that Fotolog’s growth was artificially limited due to hardware issues. Fotolog had major scaling problems with their service, and actually had to limit signups at times. They’ve now taken some funding so they can scale the hardware side more effectively. If Fotolog’s growth was indeed slow and steady, it’s may be due to the restrictions they put on signups.
At the risk of self-promotion, here’s some background on Fotolog:
http://mashable.com/2006/08/10/fotolog-4-million-users-and-counting/
besim and ethan, thanks for the good words.
pete…excellent link. great to have the backstory on fotolog. thanks!
Chris
When I started an earlier career as an Operations Research Analyst at British Airways, (some 20 years ago), I was asked to prepare a short talk on how statistical graphics can be wittingly or unwittingly misleading. I used some great examples from two fun-packed paperbacks by Darrell Huff: “How to Lie with Statistics” and “How to Take a Chance”.
I see that both are still available on Amazon. Maybe they should be more widely read.
Graham Hill
graham, agreed…”how to lie with statistics” is a must-read. will definitely check out “how to take a chance” as well.