Tags: last.fm, tweeklyfm, twitter
So, there has been interesting developments lately in the Tweekly.fm stable. Stable is actually the wrong word, because it is exactly what it is not.
It has been growing way too fast to keep it contained. It now has 13 000 users. My hosting company has been on my back for quite some time now, because I’m overloading the servers each time the tweets send. Thanks to the helpful @igitur, I have a pretty solid way of optimising the code/script. Unfortunately, as it is primarily the process of fetching data from last.fm that is slow, it is a losing battle. Inevitable down the line, I’ll keep running into optimisation problems.
I am not entirely sure where to go with Tweekly.fm at the moment. If I continue the service, I’ll have to upgrade, get its own server, get advertisers, et al in order to keep it running as it is running now. The problem with this method is that 1) I’m a full-time student. I don’t have a lot of time to spend on Tweekly.fm (which is why I haven’t actually optimised the script at all) and 2) Tweekly.fm is a mashup after all. I didn’t start it in the beginning to make money from it. It was purely a service I wanted, and thought other people would maybe enjoy. Monetising a mashup is a very iffy thing. I’m totally at mercy of Twitter and Last.fm (especially last.fm).
At the end of the day, I’m providing a service based off of last.fm. They could easily implement the Tweekly.fm feature and then I have to shut down everything. I don’t own any of the data, etc. I also don’t really have any right to this “IP” if you understand what I mean. I merely provide the conduit for the flow of data.
Which brings me to the question: do I want this? I don’t want to make a full fledged enterprise mashup. Should I go ahead with monetising and upgrading everything? Is there a way I can keep it “small” without incurring too much server problems?
Should I change the way Tweekly.fm works (less bandwidth)? There are still more social-twitter side features I can add (which I want to). Should I perhaps move it to manual tweeting? and then pay for automatic tweeting? Or spread it out across the week (user can choose when to send it)?
All these questions are plaguing me, and I don’t know where to go. After my university exams (starting next week), I’ll probably have to switch Tweekly.fm temporarily off, think about what I want to do with it, fix whatever needs fixing and take it from there.
What do you say?
Tags: tweeklyfm
I am seeing more and more people complaining that their Tweekly.fm isn’t working only to see their profiles updated later. A trend I see is that people revoke access, because they think their Tweekly.fm isn’t working, but it was working, the people only had to wait.
The problem with Tweekly.fm at the moment is that it has become slightly too big for what it was in the beginning. It has almost 11 000 users now and the script is taking about 5 hours to send. That means that tweets are spread out, by the order in which you signed up, over 5 hours. So you might think that your Tweekly.fm isn’t working, but you should just be patient… for now.
I’m currently in the process of speeding up the script. If push comes to shove, I can always optimise the script, parallelize it, add more computing power, but in the end, it’s actually a “broken” system. Down the line, it will eventually become too much. If I went all “start-up” on this, I could perhaps do it, but Tweekly.fm is a hobby project that I must manage.
So, I’m thinking of ways of changing the way Tweekly.fm works. With the new redesign, it should work perfectly. My main idea is making Tweekly.fm “manual”. Click a button and a tweet will be sent, just as it is now. If you want to auto-tweet, you must make a donation and then be able stipulate at what time your tweekly.fm tweet should run. This way, I can spread out the tweets and gather more money to pay for faster servers.
For now, I think it is a win-win situation.
Tell me what you think.
P.S. Advertising could also pay for the servers. 400k pageviews. 22k unique visitors. 400k alexa. You know you want to.