March 21, 2006
This article reprinted from the the TypePad Hacks Weblog. The original article can be found online:
© 2008, John T Unger
Attention Span Warning: This is a long post. Executive summary: a fully documented service record not only provides users with access to a company's track record, it provides the company the ability to demonstrate that outages are an anomaly and that service, communication and response are ongoing priorities.
At the time of this writing, Blogger is (now was) suffering through a partial outage due to maintenance on one of their Blog*Spot servers. This explains why I've had trouble in the last week visiting full entries for some of the blogs in my RSS feed. In the last week, I followed a number of links to archived pages on Blogger's servers that didn't load and displayed what appeared to be permanent errors. Sadly, I won't be visiting those pages when they are restored, because I just don't remember now what they were. I clicked the link, it was broken, I reloaded the page, backtracked to the index URL, and moved on when it didn't come up. That's kind of a bummer for everyone, but it's part of life really. Servers melt. It happens to everyone. They get fixed and life (or blogging, anyway) goes on.
I can't remember who pointed me to the Blogger status page, but it made me feel a little better to know what the problem was, how it was being addressed and why I was unable to click through from feeds to full posts. Now that I know those pages will be restored, I'll bookmark the permalink URLs and check them later.
Having just looked at the status page for another blog service puts me in mind to write the extended analysis of why I think the Six Apart Status Weblog needs a permanent archive. The list item on the Agenda reads:
Archived posts at http://status.sixapart.com, so users can see the history of TypePad Service and what was done to rectify outages. This is crucial to transparency. Service outage posts should not be deleted when service is restored.
Before we go any further: go take a look at the Blogger Status page and then compare it to the Six Apart Status page. Open them in tabs so you can flip back and forth a few times. Done? Great, read on.
My first impression of Blogger's status page was pretty favorable. The page displayed entries dating back to the beginning of the issue, giving me a timeline and history of the problem as well as steps taken to resolve it. The amount and quality of information was useful. Some entries had clearly marked updates showing partial success. At first, I was under the impression that Blogger was in fact offering a full archive of their status page, but upon searching for links to older entries it became apparent that this is not quite the case. The entries do have permalinks, and can be accessed after they've dropped off the Status index page, but only if you know the URL for the permalink. If you find a link to it, have it bookmarked or click through from a feed subscription, the entry is still there. It's close.
By comparison, Six Apart's Status blog is pretty spare. It displays the status for today only, no archive, no permalinks. When it does list an issue, there is rarely much explanation of what is being done to resolve the problem beyond an assurance that the problem is temporary. The feed for the page delivers today's status when you subscribe, but no older entries. In essence, it's a snapshot of the moment. Or a reflection in a passing window—once it's gone, it's gone.
So why should Six Apart change it? What possible reason would they have to provide a full archive accessible to anyone?
There are some obvious Web 2.0 answers, reeking of buzzwords: transparency, trust-building, authenticity, and so on. To be honest, when I began writing this post, that was all I really had in mind. I thought it would be useful for users to be able to really see the service record of the company over time. It was an entirely selfish idea.
But that's just the obvious reason… there's a better reason to offer a full archive, and it benefits Six Apart as much or more than it benefits users. A full archive of posts, including days when nothing goes wrong, would be good for everyone. Sure, it would be boring as hell to scroll through, since most days entries would say only "service is up." But that's exactly the point. I just didn't see it right away.
You see, people only remember what impacts them. At the end of the day, what stands out for you? Do you remember best the quiet but productive time spent working as usual, or the adrenaline rush that comes with closing a deal or arguing with a co-worker? Exactly. Things that don't change, don't stick. The status quo is invisible until something upsets it for better or worse.
When accelerated growth caused substantial downtime on TypePad blogs last year, there was plenty of talk about it in the blogosphere. On days when TypePad blogs were down, the only talk visible was from bloggers using other hosting services. Much of it was pretty unkind, and actually kind of unfair. Trouble with servers is a fact of life. It happens to everyone sooner or later. Other services went down in the same period of time… del.icio.us went down the same week, technorati has had ongoing problems managing bandwidth (and ongoing PR issues as a result).
It's like getting Slashdotted— half the time when your service is unavailable it's because everyone wants a piece of it today. But that's not necessarily going to make users feel better when it happens to them.
Kilroy vs Kaiser Wilhem: Winners write the history books, but graffiti is ubiquitous
What I want to get across by contrasting Kilroy and Kaiser Wilhelm is that people often accept at face value something which is ubiquitous. I remember the "Kilroy was here" cartoon because I've seen it a million times. But I can never remember much about Kaiser Wilhelm. He didn't make it into Mad Magazine. Blogs are popular culture— they're the graffiti of the web. If a story really takes off, the impact of seeing it in a million headlines has weight. Not everyone will take the time to find the other side of the story, or do much research on the underlying facts. They just reblog it.
The nature of the blogosphere is that it doesn't matter if you're
right or wrong; it doesn't matter if you've put together a clear and
effective message; it doesn't matter how big you are. You're
outnumbered. Unless peopel want to tell your story, you're going to be drowned out. And sadly, people prefer drama to diplomacy.
Six Apart could say they grew too fast due to the popularity of a great platform. They could say they were working on it. They could offer their users compensation. They could fix the problems. They did and said all of these things. What they couldn't do was point to a record of solid service documented over a long period of time. I think it would have helped. I think that if I had gone to the status page to see why my blog was down, seeing all those entries that said nothing more than "service is up" would have provided a calming reminder that the status quo was good. It would have been easier to say to myself, "okay, it sucks today, but it's a total anomaly." And the more information I was presented with about what was being done, the more reassured I would have felt.
Mena Trott and Barak Berkowitz did a great job of writing a series
of posts that explained what TypePad was going through, what they were
doing about it and what they were going to do about it. At
least several of these six posts were linked to from the status page at
the time, if I remember correctly. But of course, I can't verify that because there's no archive.
Here are the posts I remember:
Mena Trott: The Ups & Downs of a Successful Service
Barak Berkowitz: To our customers
Mena Trott: Update on TypePad Performance
Mena Trott: TypePad Update & Lessons Learned
Barak Berkowitz: Update on and compensation for TypePad performance
Barak Berkowitz: Compensation for TypePad Outage
Like a lot of TypePad users, I was inconvenienced and angry at the time: I had just launched a new project and gotten some pretty juicy links in, but I knew that if the page didn't load, the traffic wouldn't return. It was a limited window of opportunity and a rock had just sailed through it.
Let's look at a snapshot of what was being said at the time:
David Parmet writes a post about the service outage and other issues, with links to further commentary.
I leave a comment written on a bad day, before I finished my coffee, while service is down.
Hugh injects some reason with his comment.
Time passes, service is restored, I return and leave another comment after regaining some perspective.
A lot of people never bothered to post a second comment. And although Hugh was right that "shit happens," he was optimistic to think that it was entirely a storm in a teacup. Although it's been months since the outage in question, people still remember it far better than they remember the days when everything just worked and they were happy. Look at this recent conversation at Neville Hobson's blog:
In a post about TypePad Hacks, Neville writes:
Maybe if it all goes well, we’ll have fewer people writing about having left Typepad (though I can easily see how some of the features you have now, here, would inspire you to switch). You’ve done a great job here with layout and functionality! Nice work.
Many of the items in your list of 40 would be terrific to have at TypePad. Would they have kept me there with my primary blog? Probably not. Although I’d planned my own server/blog setup in mid 2005, it was the service issues and outages late last year and early this rather than lack of features that became the final straw.
I’m sure there will be some catharsis… I’m hoping everyone makes an effort to be civil, which has so far been overwhelmingly the case. In fact, it may be the most diplomatic comment thread I’ve ever seen on a hot topic in the blogosphere. cool.
The outages *were* a bad time. I though about moving then too, but decided to stick. I was really upset about having lost not one but a few posts when I clicked save during unscheduled outages. On the other hand, it does seem to happen to everyone. technorati’s servers are almost always under strain. Del.icio.us went down the same week as TypePad’s last big blackout. etc.
Re civility, that’s a good perspective. If TypePad Hacks manages to be a place without the trollers and those with their own selfish axes to grind, that would be something. It would certainly play to Mena Trott’s call for civility in the blogosphere.
Even though I’m here on WordPress now, I still have my TypePad blogs with my moblog being the current active one. So I’m going to participate where I can in your initiative. Not sure how yet but I’ll be in there somehow.
Thinking about TypePad for my next blog (possibly to switch)…Watching and thanks to John T Unger!
Why did I republish so much of the page, rather than just linking to it? Because I think that the thread shows a really interesting trajectory. In the beginning, Neville says he's left TypePad but still keeps a moblog there. It's not about features, it's about the service level. But by the end of the conversation, he reiterates that he still has some of his work hosted there and there's a comment from someone who is considering switching.
Would an accesible service record have changed Neville's mind? I don't think so. But I do think it would have been one more powerful tool that Typepad could have used during the crisis to show that what they had to offer on the whole was much greater than the temporary inconvenience that resulted from their growth. At any successful service, the good outweighs the bad. Documenting it and making it easy to access is a net win in more than one sense of the phrase.
More Like This: Agenda: Analysis