So last week, after a couple of months of preparation, I rolled out this new blog and published the very first post… and then I emailed my list to tell everyone about it.
And the server crashed.
Which was really, really lame… but on one level it was sorta funny – because there’s this “conspiracy theory” out there about people intentionally crashing their servers during product launches.
The only problem with that theory is that it’s completely insane. It would be an idiotic move.
OK, so once and for all… just in case there is any confusion – NO ONE who has a clue wants their server to go down when they launch. It’s a horror show and an embarrassment when it happens.
The whole point of a launch is to do your best to focus all buying emotion onto your launch day. You don’t want to frustrate people with a “Server Not Found” message at that moment.
There are all kinds of great analogies here… but just think about spending a lot of time getting a romantic partner ready for a special, romantic time, and then at the last second saying “never mind“… not good. Same thing with a product launch.
With my PLF 2.0 launch, my server crashed within 30 seconds of opening the shopping cart. The server was down for about 36 minutes.
That was an agonizing 36 minutes… and they cost me $20-30,000 in customer service costs. Those are hard costs I can quantify.
The costs I can’t quantify are the lost sales because I frustrated people… I have no idea what that was, but I’m willing to bet it was at least $100,000 in sales. Maybe a lot more. I’m sure it also cost me a fortune in goodwill.
Not to mention what this does to your affiliates… it’s pretty embarrassing to send an email to affiliates saying “Mail, mail, mail NOW” and then 15 minutes later to send an email that says “Don’t mail, don’t mail! We’re having server problems”… not good. Really, really, really not good.
Trust me, you don’t want your server to crash when you launch. It’s a disaster.
Now there’s a second thing… and that’s the idea of “pretending” that your server crashed – and then using that as an excuse to send an email to your list telling them that the server crashed.
The theory is that it gives the marketer a reason to email the list a second time, and conceivably builds some social proof about demand for an offer.
I’m sure this happens. I have no idea how often, but I’m sure it happens. Personally, I think it’s really lame. It’s a lie, and I don’t think good things happen when you lie to your list.
And besides, it’s just way too easy to come up with other reasons to mail your list. No reason to make up a lie about a server crash.
In any case, our blog crash last week wasn’t even for a real product launch… I was just launching this blog.
It was really nuts – the server was just overwhelmed by the traffic. That really shouldn’t have happened – the traffic came from me emailing my list… but this is the deal, the traffic really shouldn’t have been enough to take down the server.
And since the traffic was being driven from my email, it was somewhat “muted”… in other words, it wasn’t like I said “we’re going live at noon” and everyone was sitting at the site hitting the refresh button.
Now THAT is hard on a server.
But the traffic from an email is more gradual… and the spike will come on gradually over about the first 30 minutes after you mail.
In any case, our hosting company (Liquidweb’s Storm On Demand) enabled an “advanced firewall” to stop ALL the traffic while they rebuilt everything from scratch… on a bigger, more powerful server.
(Lucky for us… that process only took about 20 minutes, because we were theoretically using a “cloud” hosting service… unfortunately, the Storm On Demand service doesn’t scale automatically… at least as of now.)
In the end, once the dust settled, we decided to move everything over to a new host (once we realized the situation, that decision took about three seconds)… and our new host seems to have a more robust capability of scaling immediately whenever needed.
We finished that move yesterday… which means I’m free to put up a new post about it today. 🙂
So that’s the deal with the server… and I hope I cleared a few things up. Bottom line, you don’t want your server going down. And if you run into someone who acts like they know what they’re talking about – and they suggest it would be good “social proof” if your server crashed… well all that means is that you’re talking to a marketing moron.
OK, this post was originally going to be about the AMAZING reaction to my last blog entry — Powered By Internet Marketing – I was thrilled to see so many people share their thoughts… but I decided I needed to cover this whole “server crash” thing because there’s SO much mis-information about it.
So I’m going to hold off until my next post to give my reaction to the awesome conversation in that last blog post… and then there’s that little detail about announcing the winner of that Flip video camera – I’ll do that in my next post as well.
So that’s it for now. Just remember, you do NOT want any server crashes. They just plain suck.
P.S. I would love to hear your take on all this… the fabled “server crash” controversy, as well as all the other myths that fly around about product launches. Some of the stuff I hear is just plain weird… let me know your thoughts in the comments section…