Great post describing what really happened at BarcampFCT, but I have to disagree (again) with your statement:
"On the other hand, if you're creating a developer platform you'd better go out of your way to have a pristine uptime record because your users really do care about whether your service is up and running or not."
This is a fallacy: why users only care about uptime if you're creating a platform and not otherwise? You make the reader believe that you only need to care about uptime when you're building a platform. That's just not true.
Let me show you why, without using physical examples.
Example 1:
Suppose Google (a search engine -- a product, right?) is down for, let's say, 2 days straight. Which search engine will you use? Will you get upset for not being able to use Google on those 2 days?
Example 2:
Now it's flickr (a photo sharing service, right?) that goes down for 24 hours, and by accident exactly on the day that you're on a conference and need to share some pictures with your flickr contacts. You're a flickr pro user who paid ($24.95/yr) to use the service. Will you now get angry at flickr? How will you share those pictures?
Example 3:
Let's increase the cost of downtime. You're a salesforce.com user and your company depends on this service to make business. Every hour of downtime will cost you real money. salesforce.com costs $20/mo. Let's pretend that salesforce.com goes down for 12 hours, by accident on a business day. How will you feel about it? Will your life go on as usual?
So, to make it clear: your argument is based on your single experience. Downtime is as important as the value derived from the product/service/platform you're being deprived of.
Just don't use fallacies or false arguments. They don't stick.
cpinto
· 1 year ago
Hi Bruno,
you presented a strong case, maybe I should have worded my opinions a bit better. It's not that you should not care about uptime, you should. Availability of your service plays a very big role in the whole experience. But that doesn't necessarily mean that if you have a hiccup now and then you'll start losing people left and right to your competitors. But again, that depends a lot on what service you are providing. Yes, if salesforce.com is down you'll get angry mobs with torches and pitchforks knocking down your doors asking for refunds. That's because people depend on salesforce.com to make money, they're not using it for fun. If google.com becomes unavailable you'll go back to another provider but as soon as it's up you'll start using Google again, because the quality of their search results is, so far, unmatched. If Flickr is down, yes a professional user will complaint, possibly take their money elsewhere (smugmug.com for example) but probably just send the pictures via email and sort things out with Flickr.
At the end of the day, it's exactly as you wrote: uptime relevance is proportional to the cost of downtime. The bigger the downtime cost the more you'll care about it. If to that you add up the unavailability of a substitute then downtime cost increases exponentially. Essentially, with salesforce.com there is no immediate substitute, whereas with Google you go use something else and with Flickr you go back to email. (Did that make any sense? I'm not asking if you agree with me, just if it makes any sense hehe)
Bruno Pedro
· 1 year ago
It makes all the sense in the world as you write it now. Not the way you were presenting it.
Downtime value is proportional to the value derived from the product/service/platform you're being deprived of.
All Comments
Great post describing what really happened at BarcampFCT, but I have to disagree (again) with your statement:
"On the other hand, if you're creating a developer platform you'd better go out of your way to have a pristine uptime record because your users really do care about whether your service is up and running or not."
This is a fallacy: why users only care about uptime if you're creating a platform and not otherwise? You make the reader believe that you only need to care about uptime when you're building a platform. That's just not true.
Let me show you why, without using physical examples.
Example 1:
Suppose Google (a search engine -- a product, right?) is down for, let's say, 2 days straight. Which search engine will you use? Will you get upset for not being able to use Google on those 2 days?
Example 2:
Now it's flickr (a photo sharing service, right?) that goes down for 24 hours, and by accident exactly on the day that you're on a conference and need to share some pictures with your flickr contacts. You're a flickr pro user who paid ($24.95/yr) to use the service. Will you now get angry at flickr? How will you share those pictures?
Example 3:
Let's increase the cost of downtime. You're a salesforce.com user and your company depends on this service to make business. Every hour of downtime will cost you real money. salesforce.com costs $20/mo. Let's pretend that salesforce.com goes down for 12 hours, by accident on a business day. How will you feel about it? Will your life go on as usual?
So, to make it clear: your argument is based on your single experience. Downtime is as important as the value derived from the product/service/platform you're being deprived of.
Just don't use fallacies or false arguments. They don't stick.
you presented a strong case, maybe I should have worded my opinions a bit better. It's not that you should not care about uptime, you should. Availability of your service plays a very big role in the whole experience. But that doesn't necessarily mean that if you have a hiccup now and then you'll start losing people left and right to your competitors. But again, that depends a lot on what service you are providing. Yes, if salesforce.com is down you'll get angry mobs with torches and pitchforks knocking down your doors asking for refunds. That's because people depend on salesforce.com to make money, they're not using it for fun. If google.com becomes unavailable you'll go back to another provider but as soon as it's up you'll start using Google again, because the quality of their search results is, so far, unmatched. If Flickr is down, yes a professional user will complaint, possibly take their money elsewhere (smugmug.com for example) but probably just send the pictures via email and sort things out with Flickr.
At the end of the day, it's exactly as you wrote: uptime relevance is proportional to the cost of downtime. The bigger the downtime cost the more you'll care about it. If to that you add up the unavailability of a substitute then downtime cost increases exponentially. Essentially, with salesforce.com there is no immediate substitute, whereas with Google you go use something else and with Flickr you go back to email. (Did that make any sense? I'm not asking if you agree with me, just if it makes any sense hehe)
Downtime value is proportional to the value derived from the product/service/platform you're being deprived of.