Amdall Gallery

It’s a little hard to believe, but it’s been a year now since I rebooted Amdall Gallery and really got rolling on this artwork and analysis blog. Actually, it has been 13 months…I’m a little late in writing this post, as the one year mark would have been August 6th. Interestingly enough, my initial brief run with the site in 2011 stopped in August seven years ago. Anyway, I thought it might be fun to crunch some data on interactions (visitor likes, comments, views), traffic, and other interesting statistics at the one year mark. I shared something similar back in April; this will be along those lines, but with a bit extra.

Also, just a quick detour – I should mention the winner of our last custom portrait giveaway! Congratulations to Isaac A. from Ohio, who won the August giveaway through one of his daily bonus entries. That daily bonus was a new addition, so it’s interesting to see that it paid off for the winner. Isaac and I have been in contact already, and it sounds like he’s got his portrait in mind already. Once he gets me a photo or something to work from, I’ll get started on it.

Now, back to the site statistics. As with the last time I tried to analyze Amdall Gallery’s traffic and interaction data, a challenge was WordPress’s somewhat limited statistical information. The WordPress dashboard gives you some basics, but doesn’t provide a full accounting of everything you might wonder about your site. I wanted to get more in-depth, so used my old pal Excel. I started by copying the full list of comments and posts from WP-Admin, highlighting and copying “likes” from each individual post, and then pasting each into their own Excel tabs. Then I used functions to work the data into structured columns, and finally used those columns to create pivot tables and charts.

I’ll share some interesting factoids, charts, and end with tables sorted by the most popular posts:

Interactions by Post Type

• Art posts are still the most popular in terms of views, but all other types (research, tech, other) have caught up by 2% and are almost even now (51% art versus 49% other).
• The vast majority of likes are still for art posts (80%), but the other types have made some headway. In April, 91% of likes were related to art. It does make sense that most likes are related to artwork, because I believe likes only come from WordPress accounts. Artists who come across my site tend to come from WordPress, since they probably browse tags similar to mine.
• Most comments are still on artwork posts.
• I am responsible for only 42% of the comments on Amdall Gallery (370 of 877 total). Because my goal is to respond to every comment, I anticipated close to 50%…so this was a surprising decrease from 49% in April. That 42% tells me that I might have forgotten to respond to some comments. It’s also interesting that the total comments increased so much in 4-5 months, up from 356 in April to 877 now.
• So, overall: a lot of people view the analysis posts, but many more people interact with the art posts through likes and comments.

Interactions Grouped by User Types

• Adding up all total interactions (comments and likes), 34% are from a core group of 21 people who commented or liked posts 20+ times. You could possibly consider these 21 users to be Amdall Gallery’s core audience, since they engage with posts most frequently.
• 59% of all interactions come from a group of 98 people who commented or liked posts more than five times. That’s a really solid group of repeat visitors, and I’m glad to see there is an audience that does come back for more.
• 9% of all interactions are from one-off visitors, which is a decrease from 11% in April. It’s good to see that decrease even slightly, because it’s an improvement in engagement.
• 14% of interactions have actually been from me, all of which were comments (usually replies).

Visitors Grouped by User Type

• This is a slightly different measurement. Whereas the pie chart showed how the sum total of all interactions are distributed, this one focuses on users/visitors with certain numbers of interactions. It seems similar at first glance, but it’s actually quite different (measuring actions on the site versus actual people).
• 48% of users (152 people) who visited only have one interaction, while 52% have two or more. 11% of users who have visited have 10+ interactions.

Traffic Referrers

• Referral sites have a pretty clear top five including the WordPress Reader (hurrah), Reddit, StumbleUpon (RIP), Facebook, and Bing (yes really).
• The WP Reader is a personal favorite, so it’s always good to see.
• Reddit is a great example of how toxic the internet can be, so I don’t share much there anymore but still get residual traffic.
• Facebook is a new addition, I think mostly due to my Luka Doncic post getting shared in his home country.
• StumbleUpon is a sad addition to my top referrals, because it was a great site and it’s unfortunately gone now. Its replacement, called “Mix”, is pretty clunky so far.

Post Rankings by Weighted Score

• I calculated “weighted scores” for posts by taking the highest values for each interaction (likes, comments, views), determining a multiplier that would make them equal, then applying the multipliers to all interaction values.
• The weighted scores were used to determine the most popular posts with all three interaction types accounted for.
• The above scatter and bar charts list the top 30-40 posts ranked by weighted score. The first scatter plots likes and comments, with top scores being in the upper right and lower scores in the lower left. The bar chart just shows weighted scores for the top 30. The second scatter is the same as the first, but showing views and likes.
• Below, I’ve including a pivot of all the Amdall Gallery posts to this point! These are sorted by weighted score, and also show views, likes, and comments. Interestingly,  two of the three most popular posts (by weighted score) have been portrait giveaways!

Well, that’s all for this data analysis extravaganza! This was extremely interesting to me, but I have a feeling I might be the only one. It is my site after all…a random visitor probably isn’t too fascinated by trend analysis for artwork posts. But if someone read this far, thanks for sticking with it! If you have any questions about how I did this, particularly if you run your own WordPress site, definitely feel free to reach out in the comments section.

And if anyone did actually read all the way to the end, here are some random .gifs to show my appreciation:

(Unfortunately, I can’t give credit on any of these because I found them sitting in an old “funny .gif response” folder on this computer. If there’s no watermark, I have no clue where they came from. Based on the file date these were like 6 years ago, which is an eternity in meme years. Sorry originators.)