Page 2 of 4

Twitter traffic might not be what it seems

Are you using bit.ly stats to measure interest in the links you post on twitter?

I’ve been hearing for a while about people claiming to get the majority of their traffic originating from twitter these days.

Now, I’ve been playing with the twitter ruby gem recently, doing various experiments which I’ll not go into detail here because they could be regarded as spamming… if I’d conduct them on a large scale, that is.
It’s scary to see people actually engaging with @replies crafted with some regular expressions and eliza-like trickery on status updates found using the twitter api. I’m wondering how Twitter is going to contain the coming spam-flood.

When posting links I used bit.ly as url shortener, since this one seems to be the de-facto standard on twitter. A nice thing about bit.ly is that it shows some basic stats about the redirects it performs for your shortened links.

To my surprise, most links posted almost immediately resulted in several visitors. Now, seeing that I was posting the links together with some information concerning what the link is about, I concluded that the people who were actually clicking the links should be very targeted visitors.
This felt a bit like free adwords, and I suddenly started to understand why everyone was raving about getting traffic from twitter.

How wrong I was! (and I think several 1000 online marketers with me)

On the destination site I used a traffic logging solution that works by including a little javascript snippet in your pages. It seemed that somehow all visitors disappeared after the bit.ly redirect and before getting to the site, because I was hardly seeing any visitors there. So I started investigating what was happening: by looking at the logfiles of the destination site, and by making my own ‘shortened’ urls by doing redirects using a very short domain name I own. This way, I could check the apache access_log before the redirects.

Most user agents turned out to be bots without a doubt. Here’s an excerpt of user-agents awk’ed from apache’s access_log for a time period of about one hour, right after posting some links:

AideRSS 2.0 (postrank.com)
Java/1.6.0_13
Java/1.6.0_14
libwww-perl/5.816
MLBot (www.metadatalabs.com/mlbot)
Mozilla/4.0 (compatible;MSIE 5.01; Windows -NT 5.0 - real-url.org)
Mozilla/5.0 (compatible; Twitturls; +http://twitturls.com)
Mozilla/5.0 (compatible; Viralheat Bot/1.0; +http://www.viralheat.com/)
Mozilla/5.0 (Danger hiptop 4.6; U; rv:1.7.12) Gecko/20050920
Mozilla/5.0 (X11; U; Linux i686; en-us; rv:1.9.0.2) Gecko/2008092313 Ubuntu/9.04 (jaunty) Firefox/3.5
OpenCalaisSemanticProxy
PycURL/7.18.2
PycURL/7.19.3
Python-urllib/1.17
Twingly Recon
twitmatic
Twitturly / v0.6
Wget/1.10.2 (Red Hat modified)
Wget/1.11.1 (Red Hat modified)

Of the few user-agents that seem ‘real’ at first, half are originating from an ip-address used by Amazon EC2. And I doubt people are setting op proxies on there.

Oh yeah, Googlebot (the real deal, from a legit google owned address) is sucking up posted links like fresh oysters.
I guess google is trying to make sure in advance to never be beaten by twitter in the ‘realtime search’ department. Actually, I think it’d be almost stupid NOT to post any new pages/posts/websites on Twitter, it must be one of the fastest ways to get a Googlebot visit.

Same experiment with a real, established twitter account

Now, because I was posting the url’s either as ‘status’ messages or directed @people, on a test-account with hardly any (human) followers, I checked again using the twitter accounts from a commercial site I’m involved with. These accounts all have between 500 and 1000 targeted (I think) followers. I checked the destination access_logs and also added ‘my’ redirect after the bit.ly redirect: same results, although seemingly a bit higher real visitor/bot ratio.

Btw: one of these account was ‘punished’ with a 1 week lock recently because the same (1 one!) status update was sent that was sent right before using another account. They got an email explaining the lock because the account didn’t act according to their TOS. I can’t find anything in their TOS about it, can you?
I don’t think Twitter is on the right track punishing a legit account, knowing the trickery I had been doing with it’s api went totally unpunished. I might be wrong though, I often am.

On the other hand: this commercial site reported targeted traffic and actual signups from visitors coming from Twitter. The ones that are really real visitors are also very targeted. I’m just not sure if the amount of work involved could hold up against an adwords campaign.

Reposting the same link over and over again helps

On thing I noticed: It helps to keep on reposting the same links with regular intervals.
I guess most people only look at their first page when checking out recent posts of the ones they’re following, or don’t look too far back when performing a search.

Now, this probably isn’t according to the twitter TOS. Actually, it might be spamming but no-one is obligated to follow anyone else of course.

This way, I was getting more real visitors and less bots. To my surprise (when my programmer’s hat is on) there were still repeated visits from the same bots coming from the same ip-addresses. Did they expect to find something else when visiting for a 2nd or 3rd time? (actually,this gave me an idea: you can’t change a link once it’s posted, but you can change where it redirects to)
Most bots were smart enough not to follow the same link again though.

Are you successful in getting real visitors from Twitter?
Are you only relying on bit.ly to provide traffic stats?

What makes them click ?

The other day (well, actually some weeks ago while relaxing at the beach in Kos) I read ‘Neuro Web Design – What makes them click?’ by Susan Weinschenk. (http://neurowebbook.com)

The book is a fast and easy read (no unnecessary filler) and a good introduction on how your site’s visitors can be steered in the direction you want them to go.

The Obvious

The book handles some of the more known/proven techniques, like for example that ratings/testimonials of other people can help sell your product or service. Another well known technique it talks about is inducing a sense of scarcity/urgency in the visitor. Only 2 seats left! Buy now and get 33% off! It’s not because these are known techniques that they stop working.

Luckily 2/3rd of the book handles less obvious techniques, otherwise it wouldn’t be worth buying.

The Not So Obvious

A less known influencing technique is reciprocity. And then I’m not talking about swapping links with another website, but the fact that someone is more likely to do something for you after you did something for them first. The book cites some studies (I always love the facts and figures) and gives some actual examples of how to implement this in your site’s design, which is less obvious when you think about it. Want to know more ? Buy the book!

Other interesting sources

For a more general introduction to the same principles, I’d suggest ‘Yes! 50 Secrets from the Science of Persuasion’. ‘Yes!…’ cites some of the same studies (it seems there’s a rather limited pool of studies covering this subject), but of course doesn’t show how to implement these techniques in your site’s design. I read ‘Yes!…’ last year, making ‘Neuro Web Design’ just a little bit less interesting.

!!!Always make sure you’re able to measure your changes. If you haven’t yet, check out the advanced segmentation in Google Analytics (don’t be afraid because it says ‘beta’, it works just fine) and Google Website Optimizer.

Worth Buying?

Can I recommend it ? Sure, why not. I think it can be useful for anyone who ever had to think about the design or content of a site. You don’t have to be a marketing guy to want a site you’re involved with to be successful. The content/filler ratio is excellent too: you don’t need to wade through dozens of pages to filter out the interesting bits. (unlike ‘The Design of Sites’, which contains too much useless info and because it’s in dead-tree format, you can’t google it)

If you like it, you might also check out ‘Yes! 50 Secrets from the Science of Persuasion’.

Tip for people living in Europe: check Amazon UK for your book buying needs. Because of the low UK Pound exchange rate, it’s usually considerably cheaper and faster to get a book delivered to your doorstep by Amazon UK compared to having to order it at the local book store or web-shop.

Google Chrome OS

It’s about time someone took this initiative: Google Chrome OS

I especially like the following:

Speed, simplicity and security are the key aspects of Google Chrome OS. We’re designing the OS to be fast and lightweight, to start up and get you onto the web in a few seconds. The user interface is minimal to stay out of your way, and most of the user experience takes place on the web. And as we did for the Google Chrome browser, we are going back to the basics and completely redesigning the underlying security architecture of the OS so that users don’t have to deal with viruses, malware and security updates. It should just work.

I recently had the ‘pleasure’ witnessing several 60+ yr old friends and family (all respect for everyone in their 2nd or 3rd youth) buying their first pc, taking their first steps using a pc and the net.

Have you ever seen the gazillions of little ‘useful’ tools that are installed on a new standard Vista pc or laptop ? This is like learning to drive a new car and being placed in an airplane cockpit. And all the messages one gets about virus/security checks, fingerprint nog being enrolled, trial period expiring (because half those really useful tools come with a trial period), … If I was in their shoes, being confronted with this as a total newbie, I guess I’d just give up pretty soon. As a matter of fact, I actually gave up on Vista on my work-laptop, it was driving me crazy. Thank god I was allowed to install XP. I’m a Linux user at home, and Vista was such a frustrating experience that Windows XP actually felt like breath of fresh air.

And what are those people using? Email, browsing…. and maybe writing a little letter now and then or storing their photo’s if they have a digital camera. Actually (side note), I get the impression that hearing about facebook is a major motivator for the digital newbies to finally take the plunge, buy a pc and get on-line.

And OK, we’ve seen initiatives like this before, but Google is a brand everyone knows… unlike Ubuntu, Debain or Mandriva. Google = God. If I was Microsoft I’d be wetting my pants knowing Google was about to release their own OS, without a doubt fully optimized to use their own on-line office suit. On the other hand, the old adage ‘no one ever got fired for choosing Microsoft’ still holds a lot of truth.

I hope I’ll be able to give it a big thumbs up if a would-be pc-user asks me what kind of pc/OS they should go for in the near future.
On the other hand, if I’d do that, I’m pretty sure a couple of weeks later I’d get a call asking how to install this game or photo editing tool they got from one of their Windows using friends… or that nifty photo-printer they just bought. But then, I also get those questions now from newbie Windows users. It takes a couple of years before Newbie pc users understand that some things just don’t work and aren’t worth the time trying to fix them. I’d just wish they’d go back to the shop when something doesn’t work. You also don’t let you mechanic friend try to fix a problem with your brand new car. But that’s another story…

Wait and see…

Twilio – Telephony in the cloud

[repost because the original post was lost due to an error from my hosting provider]

Interested in an easy way to hook up your application to the phone system ? Check out
Twilio!

Haven’t used it yet, but it seems easy as pie!

Yes, you can probably do the same with Asterisk but figuring out the web service to Twilio is probably a lot easier/faster than installing and learning Asterisk.

Paypal encrypted button pitfalls

[repost because the original post was lost due to an error from my hosting provider]

I recently had to implement a small online payment system using an encrypted paypal button (‘after hours’). All of this in an existing modx cms installation. Modx is an absolutely handy, powerfull and totally undervalued cms system that doesn’t seem to get the attention it deserves. So there – my free mention.

I used this article and script from a company called Stellar Web Solutions as a starting point for encrypting the button data. (no need to reinvent the wheel)

Now, to keep others from falling into the same pitfalls as I did…

Bad OpenSSL Version

If you get the error:

The email address for the business is not present in the encrypted blob. Please contact your merchant.

Then you probably have OpenSSL version 0.9.8b which has a known issue with pipes, resulting in a cut-off output. This version unfortunately seems to be quite common on shared hosting solutions, who ofcourse are very reluctant to update their ‘working’ installations.

You could either upgrade your OpenSSL version, or refrain from using pipes. When using the Stellar Web script, you can try the fix posted here, which worked for me.

If you are not using the bad OpenSSL version, I’d suggest that at least something is wrong with the encrypted data.

Use the correct certificate

Don’t mix the sandbox with the live certificate, they are different!

Only use 2 decimals for values

I got the following error:

The link you have used to enter the PayPal system contains an incorrectly formatted item amount.

Turned out my calculated price contained too much decimal places for the value. It should have been only 2. When googling this I also found that it might mean you used a comma instead of a point or vice versa.

For such a widespread system, too bad paypal doesn’t give any clearer error messages, or provide more info on what the above errors could mean. I found lots of posts asking about the above, but had to really dig to find some answers.

2 great GUI prototyping tools

I’ve had a special interest in GUI design, usability, hci, … for quite some time.

The last few years, when I started working on a new feature I tried to forget everything I know about programming and fully concentrate on the user interface first: who will use it, how should it integrate into their workflow, how often will they use it, etc. Just ban any data model or database thoughts, pick up pen and paper and draw the user interface.

As a programmer you tend to focus on the data first. This is the way we fit the problem into our programmer’s mind, this is also the part that might contain pitfalls and when you ‘solved’ it you’re usually pretty sure that it’s solved in a good (enough) manner (development = left side o/t brain?).

The user interface on the other hand is usually an afterthought: you know you can do it, there’s no challenge here and it gets rushed together because there’s a deadline you need to consider. For most developers, there’s also a lack of that reassuring feeling that you’re on the right track or a big sense of accomplishment when it’s finished (UI design = more right side o/t brain?).

This is obviously not the right way to go about this: for the user, the user interface IS the program. It’s all he or she sees, and even if it doesn’t look like a rushed afterthought, it often feels like one. Most UI’s still are just a way to interact with the data instead of being a handy, useful tool. They are an interface between the keyboard, mouse, monitor and data instead of being an interface between the user and the task.

Although I’m a firm believer in ‘paper prototyping’ and designing the UI with a clear mind away from the machine, I also often miss digital tools like copy/paste and being able to email my mock-up to someone right away. (but still tend to stick to pen and paper, and not only because it’s got a better carbon footprint 😉

However… these 2 quite recent tools could fill that gap:

  • Balsamiq Mockups not free but I guess you can’t afford not to buy it when working in a pro team
  • The Pencil Project which comes as a Firefox plugin and is totally free

So far, no-one I mentioned these to had heard about them. Maybe I talk to the wrong people.

Balsamiq Mockups keeps the sketchy feeling by for example not using 100% straight lines. This might seem a bit backwards but it keeps you from trying to make your design pixel-perfect and gives the impression that your design is still very open to change and suggestions from others. Although it doesn’t feel right in a programmer’s mind, it is ideal in this GUI-prototyping context.

Do I hear someone say ‘Visio’ ? It’s probably just me but the few times I tried it took me ages to design a GUI with Visio. I’ve seen my share of Visio UI mockups, and I often got the impression that they were made by people who only use MS Office products and design an interface to the data instead of a User Interface. Doesn’t mean Visio is bad though, I guess I’m just Pavloved not to like it.

Enjoy!

Ruby live regex checker

This Ruby regular expression checker/tester is a hidden gem that’s far too low in the serps: http://rubular.com/.

Ruby crc16 implementation

I needed a crc16 in order to feed some data to a dinosaur.
Google couldn’t help me this time so I rolled my own. The pre-calc table was borrowed from a python implementation.

Hope this saves someone else the hassle…


CCITT_16 = [
0x0000, 0x1021, 0x2042, 0x3063, 0x4084, 0x50A5, 0x60C6, 0x70E7,
0x8108, 0x9129, 0xA14A, 0xB16B, 0xC18C, 0xD1AD, 0xE1CE, 0xF1EF,
0x1231, 0x0210, 0x3273, 0x2252, 0x52B5, 0x4294, 0x72F7, 0x62D6,
0x9339, 0x8318, 0xB37B, 0xA35A, 0xD3BD, 0xC39C, 0xF3FF, 0xE3DE,
0x2462, 0x3443, 0x0420, 0x1401, 0x64E6, 0x74C7, 0x44A4, 0x5485,
0xA56A, 0xB54B, 0x8528, 0x9509, 0xE5EE, 0xF5CF, 0xC5AC, 0xD58D,
0x3653, 0x2672, 0x1611, 0x0630, 0x76D7, 0x66F6, 0x5695, 0x46B4,
0xB75B, 0xA77A, 0x9719, 0x8738, 0xF7DF, 0xE7FE, 0xD79D, 0xC7BC,
0x48C4, 0x58E5, 0x6886, 0x78A7, 0x0840, 0x1861, 0x2802, 0x3823,
0xC9CC, 0xD9ED, 0xE98E, 0xF9AF, 0x8948, 0x9969, 0xA90A, 0xB92B,
0x5AF5, 0x4AD4, 0x7AB7, 0x6A96, 0x1A71, 0x0A50, 0x3A33, 0x2A12,
0xDBFD, 0xCBDC, 0xFBBF, 0xEB9E, 0x9B79, 0x8B58, 0xBB3B, 0xAB1A,
0x6CA6, 0x7C87, 0x4CE4, 0x5CC5, 0x2C22, 0x3C03, 0x0C60, 0x1C41,
0xEDAE, 0xFD8F, 0xCDEC, 0xDDCD, 0xAD2A, 0xBD0B, 0x8D68, 0x9D49,
0x7E97, 0x6EB6, 0x5ED5, 0x4EF4, 0x3E13, 0x2E32, 0x1E51, 0x0E70,
0xFF9F, 0xEFBE, 0xDFDD, 0xCFFC, 0xBF1B, 0xAF3A, 0x9F59, 0x8F78,
0x9188, 0x81A9, 0xB1CA, 0xA1EB, 0xD10C, 0xC12D, 0xF14E, 0xE16F,
0x1080, 0x00A1, 0x30C2, 0x20E3, 0x5004, 0x4025, 0x7046, 0x6067,
0x83B9, 0x9398, 0xA3FB, 0xB3DA, 0xC33D, 0xD31C, 0xE37F, 0xF35E,
0x02B1, 0x1290, 0x22F3, 0x32D2, 0x4235, 0x5214, 0x6277, 0x7256,
0xB5EA, 0xA5CB, 0x95A8, 0x8589, 0xF56E, 0xE54F, 0xD52C, 0xC50D,
0x34E2, 0x24C3, 0x14A0, 0x0481, 0x7466, 0x6447, 0x5424, 0x4405,
0xA7DB, 0xB7FA, 0x8799, 0x97B8, 0xE75F, 0xF77E, 0xC71D, 0xD73C,
0x26D3, 0x36F2, 0x0691, 0x16B0, 0x6657, 0x7676, 0x4615, 0x5634,
0xD94C, 0xC96D, 0xF90E, 0xE92F, 0x99C8, 0x89E9, 0xB98A, 0xA9AB,
0x5844, 0x4865, 0x7806, 0x6827, 0x18C0, 0x08E1, 0x3882, 0x28A3,
0xCB7D, 0xDB5C, 0xEB3F, 0xFB1E, 0x8BF9, 0x9BD8, 0xABBB, 0xBB9A,
0x4A75, 0x5A54, 0x6A37, 0x7A16, 0x0AF1, 0x1AD0, 0x2AB3, 0x3A92,
0xFD2E, 0xED0F, 0xDD6C, 0xCD4D, 0xBDAA, 0xAD8B, 0x9DE8, 0x8DC9,
0x7C26, 0x6C07, 0x5C64, 0x4C45, 0x3CA2, 0x2C83, 0x1CE0, 0x0CC1,
0xEF1F, 0xFF3E, 0xCF5D, 0xDF7C, 0xAF9B, 0xBFBA, 0x8FD9, 0x9FF8,
0x6E17, 0x7E36, 0x4E55, 0x5E74, 0x2E93, 0x3EB2, 0x0ED1, 0x1EF0
]

def crc16(buf, crc=0)
buf.each_byte{|x| crc = ((crc< <8) ^ CCITT_16[(crc>>8) ^ x])&0xffff}
crc
end

mod_rails – first impressions

  • Installing was easy as pie. I was missing the Apache development headers and the installer told me which ‘yum’ command I should use to install them. Nice.
  • When you don’t have heaps of memory, remember setting RailsMaxPoolSize to a lower value to avoid swapping. I’d experiment with RailsMaxPoolSize and saturating the number of concurrent requests using ApacheBench in any case. More concurrent threads is only better until a certain point.
  • Apache/HAProxy/Mongrel was a bit slower than Apache/mod_rails, which was a bit slower than HAProxy/mongrel (without Apache). The difference was nothing to loose sleep over. RailsMaxPoolSize was set the same as the nr of mongrels.
  • While ApacheBench testing, touching tmp/restart.txt to force a restart could cause a hang. Only restart of Apache helped.
  • I would get an occasional time-out and ab stopped (heavy testing, large nr of concurrent requests). BUT this also happened with Apache/HAProxy/mongrel (NOT when using HAProxy/mongrel ‘direct’) so I’d guess this is rather an Apache issue.

Mod_rails seems to be a surprising stable and finished product, and this for a first public release! I’m anxious to check out their Ruby Enterprise Edition.

The occasional hang when touching tmp/restart.txt makes me a bit weary, but this only seems to happen under heavy ApacheBench testing with a high number of concurrent threads. The hangs of Apache under heavy load on my test system, even when no mod_rails is involved might indicate this isn’t even a mod_rails problem. I don’t see this being an immediate issue under normal circumstances.

Thank you Phusion! (I don’t like the logo though)

mod_rails

I’ve been planning on checking out nginx for a while. Mainly as an alternative for apache in my currently preferred apache/haproxy/mongrel Rails install.

I guess I’ll have to hold on to apache a little while longer and check this one first: http://www.modrails.com/

From where I’m standing this is the most interesting thing to happen to Ruby on Rails deployment since mongrel.