Alex Schultz: APIs

Cocktail season is upon us – thanksgiving and Halloween traffic comparisons for UK and US cocktails

This Friday I saw a nice spike in my Friday traffic to cocktailmaking.co.uk, the highest since Halloween and beyond that best since the summer cocktail period. I am excited the holiday cocktail season has begun.

I have included the traffic graph for my US traffic over the last two months below. One really fascinating thing for me is that Halloween is now a really big cocktail period and the fun co-incides with the weekend due to the holiday. You can also see that Thanksgiving is great for cocktails, there is a really clear midweek spike for thanksgiving which doesn’t fit the normal periodicity of my traffic (saturday and sunday high, mid-week low).

When you compare that to the UK traffic below you can clearly see no double spike around thanksgiving but everything else looking really similar (halloween is a little less pronounced)

I love looking at different countries and seeing how they vary and more importantly how similar they are. Much of my career I have talked to people who expound the differences between markets but so often with internet marketing countries are more similar than different.

Making my site faster with Google’s coming site speed changes

Given google’s recently announced plan to factor site speed into it’s search algorithm in Q1 I decided to take a look at their new site speed tool in the Google labs section of the webmaster tools interface. The results are below.

Bottom line some really quick wins, kill the delicious button on every page of my paper airplane site that is harming my load time. Replace the old google analytics code with the new google analytics asynchronous snippet and get the whole site gzip compressed (am really bummed my shared hosting doesn’t do this already, seems an easy win).

Nice tool released by Google so you should all take a look at it and at least nail the quick wins. This will be a big deal in Q1 I am sure (although that being said caffeine as a whole is gonna be interesting).

The other interesting note today is here from search engine land google is now personalizing results for everyone with a google cookie. My search has gotten better and better thanks to this but as someone who worries about SEO I often look at the non-personalized results too. Personalized and non-personalized are diverging rapidly so really understanding user behavior and making your site compelling is growing more and more important.

Long term optimizing for user experience is key to maximize success in Google (site speed, low bounce rate etc…) and social media. I think it’s great that generating traffic and user satisfaction are coming closer and closer together.

Revenue drivers (or not) in last two years

If you have a website that gets 100k’s of visitors you can make a decent supplementary income from the web. Depending on the topic area you can make a lot. In some niches that can make you a lot more than a decent supplementary income but my sites aren’t in that category (yet :) ). That being said I am lucky enough to have enough revenue and traffic to have fun data to share from time to time. So here are some fascinating points. 4 things have driven my revenue over the last 2 years:

1. Traffic (specifically unique visitors not page views)
2. CPMs (revenue per visitor in effect, I may write another series of posts about this)
3. The bubble in video revenue around revver’s funding and the subsequent collapse in that bubble
4. The economy collapsing and now the start of the recovery

The below graph shows the collapse in revenue per visitor after Sep 07 when revver CPMs started to tank

The below graph shows year on year revenue growth which you can see took an initial hit post revver changes but tanked even harder in the middle of 08. I firmly believe advertising revenue for smaller sites is a great indicatory of economic trajectory (around the last recession my cpms tanked and then the news started talking of recession the same happened here).

Overall visitors have grown through this whole period very nicely. Page views have not since I have optimized for things like video views which users seem to love but means they view less pages.

And revenue has now recovered very nicely so that November was almost equal to the video bubble peak with revver paying out huge amounts for video views. In general I think this emphasizes that your view should always be on the long term… build traffic, build audience, make your sites better and revenue will follow over a multi-year period.

So I am focusing on integrating social more and more right now into my sites and adding more video. Looking at the multi-year web trend I believe that’s where the future is so better get to it sooner rather than later. The present will take care of itself.

a little bit of a data issue with google trends for websites

I don’t think I am giving anything away when I say the facebook stats page at the start of 2009 said: 150MM users 50% of whom log in daily i.e. 75MM daily users and today says 300MM 50% of which log in daily i.e. 150MM daily visitors. Hitwise, Comscore, Compete, Nielsen etc… tend to back this up. However Google trends for websites not so much:

I think there is some more work to be done here esp. since one of the largest sites on the web can have such inaccurate data on Google trends. This is a shame since I really think google trends for websites and adplanner are awesome tools and this makes me lose a little faith in them :( That being said plain old Google trends data looks fine.