Escaping the Traffic Doldrums with an Upgraded Hosting Account

I noticed something was wrong when I clicked around on my site and noticed a lot of “Error 500 – internal server error” messages. Traffic had been pretty steady for the past few months, but as I later found out, my WordPress installation was gradually becoming more and more complex, using more server resources with every upgrade. So the humble little 8$ a month plan I’ve been using since 2012 was no longer enough to serve all the pageviews necessary for a site that over 100,000 people visit each month.

My host had subtly discouraged me in the past from upgrading to their “enterprise” plan, noting that tests comparing server response time between the basic and enterprise plans had not been done. I really appreciate how honest they were about that, but had I upgraded then I would have likely captured much more traffic in the months between then and now.

I never thought upgrading my hosting account would make such a terrific difference. This is what my Google Analytics account looks like for the period between last August and today, Feb 15 2015.

traffic-pattern-after-server-upgrade

I hit a record for most visitors in a single hour the very next day. It was partly coincidence, but without the extra resources my traffic stats would probably have remained unchanged from the previous day, since they seem to have been maxed out as it was. My site traffic over the month and a half since has been up an average of about 50%.

The hosting account I switched to was “enterprise basic” from my host, StableHost. It costs 20$/mo instead of 8$/mo and gives me 8x the processing power. Processor power is extremely important for a PHP-based site, which includes WordPress, Joomla or Drupal CMS based websites. However, a web host will rarely advertise their processor power, instead focusing on irrelevant shit like “1000 email addresses!” Or “Unlimited file transfer!”

Spring and Summer Search Boom and Bust

This is sort of a tradition now, a twice yearly look back at that most effortless river of traffic, organic search.

Last time I checked it was Feb 21. The traffic growth rate was high and uninterrupted. This time we’re introducing a bit more ambiguity.

period-2-traffic-growth

As you will note, search traffic growth took off during the summer of 2014, and tapered off toward the end of summer.

These were Period 2‘s statistics.

week of aug 18: 3115
week of feb 9: 6223
growth amount: 3108
growth per week: 124

And now, Period 3:

Week of aug 18: 3115
Week of Sep 28: 7904
Growth amount (34 weeks): 4789
Growth per week: 140

What’s up with Summer?

My most popular article by an enormous margin this summer was a summer-time only topic about keeping cool during hot days. It’s not terribly surprising to see traffic dropping off, although I’d hate for this drastic trend to continue!

But it won’t. Here is the same chart with traffic from just that summertime article overlayed in orange. Search traffic from the “keeping cool” article has disappeared almost completely, from its high of over 6,100 per week, down to last week’s 391.

trafic growth overlayed summertime

I’ve always been surprised that my sites never get that consistent weekly shape that most sites do. Maybe instead I’ll have a consistent seasonal shape.

 

Buying Websites as an Investment Vehicle

I joined a Mr Money Mustache Meetup group. MMMMeetup I’ve cleverly called it. On the hike we completed last weekend, I spoke at length about building, buying and selling websites.

lynn-creek-canyon

The investors among the group perked up when I mentioned that websites trade at a P/E ratio of 2. That’s a deal an order of magnitude better than stocks or houses, which trade at ratios of between 20 and 30.

I realise there’s a good reason that websites trade at such a high discount rate; they’re inherently much riskier than owning land, labor or capital. These are some of the risk factors I’ve identified in a conversation with one of the MMMMeetup members. 

Things that go Wrong with a Website

  • A site with traffic based entirely on search might be using shady practices to get search ranking. A google algorithm change can wipe out search traffic. This can be a death sentence for a site, which is why vetting beforehand is so important.
  • Server can get hacked and will require repair. Repair can take hours to days to fix fully. This happens fairly frequently with WordPress installs.
  • A static site might start to slowly lose relevancy and will need fresh content to remain competitive in its space.
  • A competitor site might be launched, taking a portion of your traffic.
  • A site relying on time-sensitive information will become out of date (news sites, science&technology, health fads) and will need to be updated
  • New browsing technology may require a site rebuild. The best example of this is the smartphone trend toward sites that work on any size screen. A more recent example is the move to SHTML as a more important ranking factor.
Not every sudden change is negative.
  • A direct sale advertiser wants to place sponsored posts or custom ads, doubling your revenue for one or more months but requiring an account manager.
  • Your site explodes in popularity and requires a server move and/or content delivery system.
  • A new affiliate program needs to be integrated site-wide

Where to Buy Websites

I’ve identified these 3 marketplaces as the best at this time.
  • Flippa.com – This is the biggest website exchange as far as I know. Lots of cheap sites but there’s a lot of “pump and dump” type sites for sale – buyer beware and know how to do your due diligence on a site before buying.
  • EmpireFlippers.com I don’t know that much about it but a colleague of mine says its the only one she trusts. I think this was built around a book or something.
  • Quiet Light Brokerage – premium sites, usually costing $1 million and up. Well vetted by the brokerage beforehand, and you have to sign an NDA to get on their list. Not that many sites come through this pipeline.
Craigslist has sites for sale too, although it’s not a great fit because Craigs is local focused. I’ve only ever seen one decent site for sale on craigslist.

Considerations for Vetting a Website

  • Start with using SEMrush. Its stats are much more useful than compete.com or alexa, which are popular but I wouldn’t bother with at all. This gives an indication of a website’s potential for passive income through search engines. Although the subject matter of the site will have a huge effect on its income potential – finance related sites for example can push credit cards and startup business sites can push hosting services, which both pay a positively massive commission.
  • Check the website with mozrank, open site explorer and Google Pagerank. What you’re looking for are high quality links and a decent track record. 
  • Also look into their demographics. I have 2 popular sites, one which caters to males in their 20s and another primarily used by men and women over the age of 55. The site for 20 year olds hardly pays its server expenses, the 55+ site performs amazingly.

Anatomy of a Traffic Surge from Various Social Networks

I’ll be analyzing referral traffic patterns for the three large social networks I get most of my traffic from: Facebook, Reddit and Pinterest. I don’t have much  traffic from Twitter, Digg, Stumbleupon, or BuzzFeed so you’ll have to look elsewhere to find those traffic patterns.

All graphs used cover the same exact time scale, Sep 1 2013 to Feb 24 2014.

Facebook Traffic Anatomy

My Facebook referrals deliver a massive amount of traffic which tapers off to to a small traffic trickle after a day or two. It’s hard to see on this graph, but the slow period after my spike of 10,000 visitors at the end of 2013 is about 100 visitors a day. That’s a decay of around 99%.

facebook-same-time-period

Graph represents ~10% of total traffic for this time period.

Reddit Traffic

reddit-same-time-period

 

Graph represents ~2.5% of total traffic from this time period.

Reddit traffic seems to decay at about the same rate and amount. Both Facebook and Reddit feature similarly time-sensitive news feeds which is why their traffic patterns are so similar. But Pinterest works a little bit differently…

Why I Love Pinterest

Pinterest delivers a surge of traffic that stays put which is why I’ve been focusing more and more on promoting there.

pinterest-traffic-patterns

Graph represents ~19% of total traffic for this time period.

Unlike the other social networks above, Pinterest isn’t time sensitive. A post on Pinterest propagates from user to user, so every time it gets “pinned” it’s published fresh to each new pinner’s network. Rather than taking a couple days, decay takes months and at least for me, has leveled off to a full 25% of its initial traffic spike. That’s 25 times better for long-term traffic than either Facebook or Reddit.