Petopia: The Dangers of Blind Data
So I’ve been thinking about updating the navigation menu for Petopia just a tad to make it a little more useful. The first question I had, of course, was how people use the site now. For that, I turned to Google Analytics, which records and analyzes website traffic patterns. For instance, it can tell you which pages your users enter your site through and where they go from there.
The data I got from Google Analytics was just fascinating, although some of it was a little unexpected. It looked like people were going directly to the pets and skins they wanted, and very few were passing through the family pages (which I had tended to think of as the traffic hub of the site). But hey, if that’s what people do, then that’s what they do! I took the info I had and started working on what would theoretically be a much more useful navigation bar for Petopia.
A week later, I noticed that I had forgotten to include the Google Analytics code on any of the pet family pages. No code meant no data; no data on family pages meant that my theories on how people use the site were severely warped. Indeed, I put the Google Analytics code on the family pages and immediately saw that people do use the family pages — in fact, more than I had initially expected. So it was back to the drawing board!
If I had sanity-checked the data I was getting against what I had believed to be true earlier, instead of blindly trusting that Google knew better than I did, I would have saved myself a week of development. Of course, it works the other way also: if I blindly trust that I know how people use the site, I don’t need Google Analytics at all (and I still wouldn’t know exactly how incredibly popular Humar is).
Both of these problems happen a lot in MMO game development. It costs a lot of time and effort to gather data on how people play your game, so it’s a lot easier just to assume that most people play the way you have seen people playing — the way you and your friends play. On the other hand, when you do set up a system to gather data — say, for instance, something that records the damage done by player characters and makes it into a handy graph — you’ve put enough time and effort into gathering that data that you tend to accept the results without doing nearly enough checking on their accuracy … or thinking enough about their applicability.
7 Comments
You’re cute |:) I remember looking at google analytics for the first time and thinking the same thing.
btw love ur site even if i only go on to stare are boars and raptors.
Maybe u should put a site survey on the site or something like that to get some actual user opinions.
The site’s awesome but I agree the nav can use some work, oh yeah and you should test it too. Like have ur sigificant other or someone given a task (ex. “You want a pet that does really high damage with lots of health and you’re not too worried about armor. And it has to look awesome”) then watch them as them try to complete the task without helping them.
You’ll learn a lot. gl
Hi, Im am writeing you why I first came to your web site and why I keep comeing back. First off my pet is a ghost saber named Osiris. Named after the Egyption god that would go out and punish the wicked. I came to your web site so that I can find other pets that would teach me higher pet skills so I can train Osiris. Eg: claw3 to claw4. All though now that im am here I have read and enjoyed your whole web site. I love your new calculator for pet skills. I find it interesting that WOW will allow you to go to skill level 11 in stamina. But, if you do train your pet to that level you can pretty much forget about training your pet in anything else. I also like your work on the break down of each skill and its cost effect ratio. I hope you do the same for prowl. I have found it more useful than dash. Now theres another artical for you. What skill are best for your pet. Do a servey from those that enter your web site. what is your favorite pet skill and why. Thank you for this web site and allowing me an opportunity to make these comments.
Brontosaurus: Er, thanks. :> My significant other actually wrote the database that we’re incorporating into the site, so I’m not sure he’s the best judge. But the nice thing about the work I’m doing now is that it will make Petopia easier to adjust as I get feedback on the live site. And this blog is a useful place to solicit feedback also.
J Michael: Thank you for the feedback! Both the calculator and the skill analysis you mention were written by other people — Petopia readers and hunters who just wanted to share what they’d found out — so I can’t take too much credit for them. :> I would like to get more analyses like that also. Perhaps you could do one yourself and share it with us!
Just a thought - I came to Petopia on more than one occasion to find an “instance” pet. I have Humar trained as a tanking style pet (and he rocks so hard, I can see why people look him up constantly), and so I wanted a pet that was less likely to steal tank aggro (even with growl turned off and him being on passive. Mind you, PuG tanks can often suck…)
Anyway! My point was, as good as the Offensive/Defensive/Well Rounded classifications are, for a Hunter noob like me having a few additional details like “good for instances” or “good as a tank” could be handy.
Or maybe its just me being a bit thick…
You could just ask us what pages we like :)
For example, one of the pages I miss was the one that had the table of pet families vs. food. See, I’m one of those lazy hunters, so I tend to keep pets who eat fish :) (fishing, the hunter only talent :) )
Krush
Kittyn: I agree. With scaling and the increased importance of skills, I tend to think the Offensive/Defensive/Well-rounded classifications are a lot less useful than they once were. Something that takes into consideration the available skills as well as the highest stat (which is all the old classifications really used) would be a lot more useful, I think. On the other hand, it’s also a lot fuzzier and tends to change as people try out various strategies. I mean, who would have called a scorpid a DPS pet six months ago? And a lot of people still don’t today, because it really depends on how you use him.
Krush: One of the goals of getting Petopia switched over to CSS is letting me spend more time on those pages (like diet, and popular pets) that have fallen by the wayside. I’m actually working on a slightly cleaner version of the diet by family page today. :>
Well, mainly what makes the scorpid such a DPS pet is his poisons.
Leave a Reply