How a 9 Day Denial of Service Attack Affected Blog Traffic

February 28, 2010 By 21 Comments

(Reading time: 4 – 7 minutes)

Got lemons? Make lemonade. Such a hideous old cliche, all the worse for it’s truth. Having your website squeezed will leave a sour taste, but the resulting traffic data turns out to be sweet.

Now that there’s a year’s worth of data for Website In A Weekend, let’s use the stupid service attack as an excuse to take a few notes. No fancy writing today. Just the facts, with four graphs showing traffic.

Daily stats for February 2010

WordPress daily stats for February 2010

WordPress daily stats for February 2010

Examining the graph above:

  • The red bar shows a local plateau in traffic, which is good, exactly what I was aiming at (10k hits/month).
  • February 11, 2010 disaster strikes early Thursday morning. Hosting company informs of distributed denial of service “This one is as bad as we’ve ever seen.”
  • Traffic stays cratered for a few days. Literally, cratered. Kind of cool when when think about it. I continue to post, but comments are turned off.
  • When the server was offline completely on Feb 18, I gave it up. No articles for 3 days in a row. That’s the longest posting break since May 2009.
  • Traffic starts to return to normal on February 21 when I run Julie Angelos article on being successful.

Weekly stats for last 6 months

WordPress Weekly traffic numbers week ending February 27, 2010

WordPress Weekly traffic numbers week ending February 27, 2010

The weeklies show a similar story.

  • There’s a definite crater through the holiday season. Note that my Alexa ranking, which is a relative index, remained steady or declined (good) during the holiday season. While total traffic was decreasing, I was gaining traffic share. Nice!
  • Since the attack started half way through the week, the first week of the attack doesn’t show much decrease. The second week is pretty awful. The screenshot was taken 5 days into the 3rd week since the attack began, so there is still room for growth on that number. The rate of traffic loss clearly declined in any case.
  • The traffic plateau shown for the first four weeks (i.e., January) was my target: about 10k hits/month, then twiddle the knobs to balance ROI for time invested.

Monthly stats for first year of Website In A Weekend

WordPress Weekly traffic numbers week ending February 27, 2010

WordPress Weekly traffic numbers week ending February 27, 2010

The monthly is where it’s getting really interesting. I’m giving you the whole year too, no secrets here.

  • April is an outlier, one day of bogus traffic from StumbleUpon. Take that one day away, you would see smooth growth all the way to September.
  • Starting in September (after Burning Man), I started commenting on other blogs more often. You can see the result in my traffic.
  • No surprise that December is cratered. Take the Christmas Crater out, growth would have been pretty smooth.
  • And here’s where it’s really interesting: despite being throttled and without even commenting for 9 days, February 2010 is already my second best month, and that’s with 2+ days left in the month when the screenshot was taken. As of February 27, 17:54 Pacific time, the February count is 8433, and it should hit about 8700 by midnight February 28, 2010.

Relative ranking from Alexa

Let’s see what Alexa has to say. Here’s the trailing month:

Alexa trailing 1 month traffic rank

Alexa trailing 1 month traffic rank

  • A one week decline from mid-35k to around 100k, then bouncing around at 100k. Ouch.
  • On the other hand, I do nothing and I still rank right around 100k on Alexa. That’s pretty close to money for the right readership. This is good to know.
  • Getting traffic back the second time turns out to be much easier. It took months to get from 100k to 40k. I’ll do that again in a week to 10 days.

Note: Alexa is very useful for relative ranking. It doesn’t correlate well with raw traffic, but that’s ok. The relative traffic is still useful.

Supposed you’re attacked, what then?

First: don’t sweat it. Seriously. If you’re running a solid operation, it’s not a big deal.

Here’s a few suggestions:

  • Do something else. Literally. Step away from the computer. Go outside. Find something else to do.
  • Write small articles which will load quickly. Tell your readers what’s going on, but don’t get all weepy about it. I have no way of knowing for sure, but I suspect had the server not gone down completely on February 18, the whole 9 days would just look like a more or less ordinary “slow traffic week.”
  • Have a plan you can execute very rapidly to keep your readership engaged. I have a couple of ideas I’m toying with. If it happens again, perhaps I’ll implement those.
  • Don’t sweat it. These things happen to everyone. Even Google gets hacked.

What about you?

Do you have a plan in case of attack?

What about other internet emergencies, are you prepared?

Let’s talk it over!

Filed Under: Building Traffic Tagged With: , , ,

Saturday Morning Surfing: How’s Your Blog Traffic Lately?

February 27, 2010 By 36 Comments

(Reading time: 3 – 5 minutes)

Some high traffic blogs don’t earn much money, while some blogs with much lower traffic yield a substantial amount of revenue. But in general, more traffic is better.

There’s a lot of ways to increase traffic to your blog. Let’s make it simple and sort traffic generating activity into either

  1. Passive: You don’t go get traffic, the traffic comes to you via search engines, people linking to you, etc.
  2. Active: you go get traffic by commenting on other’s blogs, actively acquiring backlinks, purchasing advertisements, etc.

Passive traffic is really cool. I get a few dozen hits per day for a handful of articles that are really popular, and I estimate about 30% of my traffic is coming in from search or links.

Most of the rest of my traffic is a result of active effort. Commenting and posting on forums really works.

But no matter how you’re generating traffic, there comes a time when your efforts are not rewarded. You reach a plateau.

So, what to do…?

How your blog traffic reaches a plateau

Let’s back up for a few minutes, and read Mike CJ’s article on The Traffic Plateau and the One Dimensional Blog. Recapping, Mike claims most bloggers go into a feedback loop where writing for a self-selected audience stymies growth. You start writing what your audience wants to read, then your traffic plateaus.

This seems reasonable.

Consider Weird Al, or the Aquabats. Both have a certain audience that resonates with them; neither are likely to produce strings of hits doing what they currently do.

They are making a living, but they have reached a plateau.

On the one hand, having nice, stable traffic with nice stable income seems desirable. On the other hand you don’t want your blog to get stale.

Here’s my take (pure S.W.A.G.): once you have an operational system for running your blog, that system will inevitably take you to a plateau, and no further. You will find working harder doesn’t pay off.

This is good.

The key is put your system in place, then work like stink until you hit refusal. Then pull back until your effort matches your return.

Sometimes, you get the experience of pulling back imposed upon you.

Lessons from denial of service attack

If you don’t already know, Website In A Weekend was under a denial of service attack from February 11 to February 20.

This resulted in Akismet being non-functional, massive spam, and really slow load times.

I turned off commenting and stopped commenting on other blogs through that period of time.

I learned my audience measures about 100k on Alexa when I simply write, and do very little promotion. That’s a “natural” traffic plateau of about 175 hits/day (WordPress Stats).

If I spend time commenting, I can drive the long run average down. My Alexa 3 month moving average is back down around 60k now, and would probably be below 50k on Alexa had I not had (seemingly) half of February subjected to a denial of service attack.

Here’s what I found out over the summer of 2009: Once you have a plateau, it’s not difficult to twist the knobs and break out of it. Back in September, I started a commenting spree which broke Website In A Weekend into the sub-100k range on Alexa.

Another interesting thing, which seems to be true, is once you “build your audience” the first time, it’s easier to build it back a second time.

My end goal is acquiring a stable traffic base here on Website In A Weekend which converts at “ramen profitable,” so I can hunker down into some serious coding. That is, I’m working towards being a “self-funded startup.” Finding a traffic plateau which supports that would be very nice indeed.

There’s a lot left to learn acquiring and maintaining traffic. I’ll write more about it in the future, and I’ll have some traffic graphs and statistics covering the denial of service attack.

How’s your traffic? Have you reached a plateau? If so, are you going to try and grow beyond it, or is the amount of your traffic serving it’s purpose for you?

Filed Under: Building Traffic Tagged With:
« Older Posts