Feel free to go and sign up now and ask any questions or respond to the feature discussion thread. It’ll take a little bit of time to get the level of activity on the site up to a good level, but in the mean time I will be personally responding to any reasonable questions posted there.

Now, however, work has been going on for a while on Traffic Android 3, and I thought it would be worth putting up a short post to let you know about it!

Previously, Traffic Android was mainly an automated bookmarking service with some tacked-on features such as RSS submission. Traffic Android “Pro”, which was the online hosted version of the product, was really just a port of the original desktop software that was written for Windows and as such, had a lot of limitations. In addition, the decline of social bookmarking and an attempt to transition it to a new architecture ended up causing a lot of problems and in the end, due to insufficient resources to properly support the customer base at the time, the service was shut down. As you might imagine, some customers were more than a bit annoyed about this, though most were understanding.

What was needed was a complete reenvisioning of the product and a few months ago I pulled out my ideas list, dusted it off and began to build a new mindmap of how a new Traffic Android would look, modelled on more up-to-date traffic-building processes yet built with the importance in mind of keeping the product super adaptable as people’s automation needs evolved with the times, and taking into account all the various feedback that had been given with the two previous versions.

What has come out of that is Traffic Android 3, which is more than a traffic-building platform; it’s now a very capable general web automation platform. I’m expecting to be running some proper automation tests this month and hoping to have a beta ready next month, so keep an eye out and make sure you’re on the mailing list.

In the next blog post, I’ll talk about how Traffic Android 3 works, how you can build your own automation products with it and how it offers a level of versatility and scale in your traffic building (and other web automation processes) that have previously been unheard of in this industry.

In the mean time, follow me on Twitter at @NathanRidley for updates on my progress. The design and resultant flexibility and stability of the new system means that version 3 is going to be around for a long time!

The new developments are:

• A brand new interface that is needed to make inclusion of new features much easier. The interface makes heavy use of JavaScript and AJAX so you’ll find it a lot more responsive than the current interface.
• A big revision to the task processing system which will ensure that updates to tasks and task processing in general will not affect your use of Traffic Android and will also allow some important additions that are planned such as automatic account creation and validating that your account profiles all work.
• The task system revision also means we’re going to be able to put in a lot of new services easily and revise the old ones more easily as well. This means lots more RSS aggregators, new social bookmarking sites, fixes to any existing tasks and new social news sites.
• Once these are done there will be some new RSS features going in, including being able to add your own feeds (other than the one we generate for you) for submission to the aggregators and also the (optional) ability to automatically create and submit campaigns using your blog’s RSS feed(s). This would mean that every time you make a new blog post, Traffic Android could automatically find out about it, pull it down and start bookmarking it for you, or at least do most of the work so all you have to do is check the settings and hit submit. More on this to come.

The Text Spinner

The feature we have in testing is “text spinning”.  It will take effect in all your project and campaign settings once testing proves it to be working as expected. Here’s how it works:

When you enter a title, description or other text, you can use special formatting to have it select random snippets of text to be used which will cause each task that executes to post slightly different content. The more snippets of text you provide, the more variation you can create for content posted to each site.

Example:

Hello, my name is {Jack|Jill|Chris|Kate} and I am {{ten|twenty} years|not|very} {old|young}.

As you can see, the selection of a random name in the snippet is fairly straightforward. The second part of the snippet is a little more complex, but demonstrates how you can embed random selections within other random selections. Some possible outputs for the above string would be:

Hello, my name is Jill and I am twenty years young.
Hello, my name is Jack and I am not old.
Hello, my name is Chris and I am ten years old.

• Tasks weren’t processing over the weekend due to a bug that has now been resolved. If you’re experiencing an extended “queued for processing” message today, give it a few hours while the backlog from the weekend clears.
• An incorrect number of profile credentials filled out was being reported on the account profiles screen. This calculation has been fixed.
• Occasionally when editing project or campaign details, some people were experiencing the wrong details coming up. This should be fixed now.

So, things are about to change with respect to how news and updates are posted. Inside the product there is actually a news and updates section, but I’m planning on making that section redundant and using this blog for all product-related updates, small and large. This means that there will be periodic micro-updates when the status of certain tasks changes, when small things are fixed, and so forth.

Also, whilst I am leading the project, as was always the case, I am now working in conjunction with a team of brilliant internet marketers in Manchester (UK), known to the general internet marketing community as “The Manchester Mob”. Partnering with these guys has opened up many opportunities in terms of both development and support resources and potential audience to whom the product can be marketed.

What’s Been Happening Recently

Up until December, the system was undergoing a private beta test which involved around fifty users. Around mid to late December, in partnership with Dan Raine, we launched an open beta to the members of The Immediate Edge, an internet marketing training membership site. This gave us a significant influx of new users and support requests to help push us forward to launch. Then Christmas and New Years hit and things slowed down temporarily, as is typical for that time of year. Things are back full steam ahead now, though.

What’s Being Worked On Right Now?

In no particular order, the following items have priority:

• Fix a bug where sometimes the information pulled from a site for autocompletion purposes is incorrect.
• Fix scripts for sites that have made updates (Digg, Propeller, Bibsonomy and Connotea)
• Fix a small bug where the number of account profiles owned is being misreported
• Ensure the subscription/billing code is tested before launch
• A few other miscellaneous very minor bugs
• Get a new website ready for launch to the public.

Once these are fixed up, we plan to do a preliminary launch to members of the Traffic Android mailing list, which you can subscribe to here if you haven’t done so already. I expect this should be within a week or so.

Following that and assuming all goes well with this launch, the big launch will likely follow shortly after that and will be performed in conjunction with Dan Raine and his crew. This will be the point at which our doors become open to the public and stay open.

What’s On The Agenda For Traffic Android Beyond This?

Again, in no particular order:

• A large overhaul to the task execution system which will dramatically increase the rate at which new tasks can be implemented. Currently creating or overhauling a task takes anywhere from twenty minutes to four hours. The new system that will be going on should shorten this to less than ten minutes per task script across the board.
• The overhaul will mean we can add a whole lot of new tasks, including RSS aggregator sites, to the mix.
• Some fancy behind the scenes work that will involve separating the task processing from the Traffic Android product so that both can be maintained separately. The technical details aren’t important here, but it will make the product more maintainable and extensible, which makes it easier for us to grow it into something much better in general.
• The ability to add your own feeds that can be submitted to RSS aggregators on top of the Traffic Android generated feed for your account.
• Some new (secret) features that will complement the existing functionality and improve the product’s ability to help you build traffic to your website.

I will post another update here when some fixes have been made. Please direct any questions to support@trafficandroid.com.

Now, onto features and technicals, for those of you who are interested.

• Proxies have been a tricky issue. Due to the nature of the web and the potential number of customers that will be coming on board, the one-private-proxy-per-customer strategy has needed to be adjusted to avoid cost issues. What we’re going to do is be running on a public pre-filtered proxy list that rotates periodically, with a list of exclusive private proxies rented as a backup measure if the public proxy list ever experiences issues. With that in place, Tor will be investigated as a potential way to truly randomize the proxies we use.
• I’ve now got all the subscription code, account settings editor and billing history page in place, which were loose ends to tie up.
• Shortly we’ll be taking the desktop product off the market for the time being, as it will largely be deprecated by the web product and we need room to get a pre-launch campaign in place. This means that you can either add yourself to the mailing list (scroll to the bottom of any page on the Traffic Android website) and wait for an announcement, or you can buy the desktop product in the next few days as a way of getting a pre-release special deal. This may be in the form of a several-month-long free trial, or a slightly different subscription price. As stated above, the details are yet to be determined.
• I’ve put in a nice error tracking system so that if the system ever throws one of those nasty errors you normally see when the programmer screws up, the entire details of the error are logged to disk so I can go through them and see exactly what happened.
• As soon as I finish up the adjusted proxy code (which will be done within a few hours of this post), my next task is to go through and tidy up any loose ends, including a small but very important task involving making the actual task processing engine a bit more robust and less prone to the occasional error that I’ve been logging during the beta process.
• Finally, we’re going to be adding a bunch more services to the release product if there’s time, and will be adding lots more services in the short term in any case.

Bottom line, development on the new version of Traffic Android is very active and we’re gearing up for an important launch will boost the product’s profile in a big way.

I have now finished integrating a live human-processed CAPTCHA service into the system. What this means is, any time a task executes that would otherwise require you to read some text from an image and type it into a box, the system grabs the image, posts it off to the service with which I’ve integrated, then within a few seconds a person on the other end, hired to sit there processing CAPTCHA images, translates the posted image and it’s posted back to Traffic Android which finishes processing the task. This means that Traffic Android Pro can now process the hardest CAPTCHA images out there in near real time with almost 100% accuracy. Pretty cool, huh?

As soon as I finished implementing this feature I updated a script for one of the sites Traffic Android is supposed to support that requires CAPTCHA input and “turned it on”. Several testers have imported their databases from their local desktop versions of Traffic Android, so this particular task has been disabled up until this point, meaning that seeing as I turned the task “on”, the task processing queue suddenly went to full steam running through processing all the new suddenly-ready-to-execute tasks that had been sitting dormat until this point.

The result? Seems to have worked reasonably well! So I’ll keep an eye on it from this point onwards and start adding new task types, such as Digg support and others. Once I have a few more tasks implemented, I have a few minor upgrades to do to the user interface based on feedback from testers and then I need to implement billing which I deliberately left until last seeing as it was the least important item to test up until this point.

In other news, one of my friends showed me another social bookmarking product available that I wasn’t aware of and it claimed the ability to post to over 120 sites! And this was a small product developed by a single programmer on the side of numerous other projects. I must say I was a little suspicious. Now, I’m not going to test that particular product out, at least not at this stage, but feedback from some of my customers has largely been that a lot of these other products are quite unreliable for the simple fact that they code them up and and leave them alone without taking into account the fact that many of the social bookmarking sites change their interfaces periodically, new sites pop up and old sites close down, which means that many social bookmarking products get out of date very quickly. One prominent social bookmarking service I was looking at the other day is still advertising posting to over 100 services, yet when I clicked through the sites they say they are supporting, half of them are no longer even in existence! So if you’re looking at buying a different product or using someone else’s service, make sure you do your due diligence first. These things require ongoing support to maintain and keep updated, which is one of the reasons Traffic Android Pro will need to charge a monthly fee.

One last thing, some people have asked if this means the desktop version will be going away. The answer is no, it will be staying, although it will require you to create a once-off user account so that the application can pull its updates and task scripts from the web, which is how the desktop product will stay up-to-date.

No matter what product you buy or what strategies you implement in order to build traffic to your website, it is absolutely critical that you don’t do so blindly. There is no quick magic bullet to building your traffic and doing so without an idea of what you are doing or why it should work can not only be ineffective, but it can be damaging to your website’s ranking.

Google employs a lot of very smart engineers and their indexing algorithms are not only hugely complex, but getting more and more intelligent as time goes on. There is a lot of money to be made by exploiting its inability to completely differentiate between human and non-human activity, but if you do so, you need to be aware that the more “obvious” your attempts to game the system, the quicker your attempts will be spotted and “dealt with” by the algorithms.

This is why Traffic Android allows you to spread your link building over a period of time. From a search engine’s perspective, if a new site appears and then a big blast of inbound links appear practically all at once and then stop, it creates somewhat of a red flag that somebody has created a new site and then quickly built some links to it to get it indexed and ranked. Human activity does not work this way. In the real world, if a new site appears and a few links appear to it, people will start to find it naturally and if the content is worthwhile, more and more links will appear over time pointing at the page or website. The difference between how human your traffic-building appears to be and how automated it appears to be, is often referred to as the “footprint” left by your campaign.

If you want to have a chance of making your activity leave as small of a footprint as possible, you need to tailor your traffic-building campaign as much as possible to appear to be the work of natural human activity. This means beyond your social bookmarking activity, you should be employing other strategies as well, such as article marketing, video distribution and anything else that is relevant within the context of your website.

So, with all of that in mind, how should social bookmarking be employed as part of the overall process? In my opinion, aside from creating naturally highly quality information that will go viral by itself, the most effective method is to build “feeder sites” that are not part of your main website. To do this, create several mini-sites that provide useful information about some micro-aspect of your main topic and make sure the site is of high enough quality that a visitor would might conceivably bookmark that site themselves. I believe that if a visitor will bookmark a page or website, then that is a good measure of a base level of quality you should aim for. In addition to mini-sites, you can also create auxiliary pages on sites such as Squidoo and Hubpages, as these will add to the overall strategy and don’t require you to provide your own hosting or resources to maintain.

Each of your mini-sites should link to your main site; not exclusively as site-wide links, but in a relevant way from within the content of your mini-sites. You can then build a social bookmarking campaign to each of these mini-sites to boost their authority to a basic level within Google’s search index, and this base level of authority will then be funnelled up to your main website. Because Google is getting better and better at detecting non-human activity, it is best to keep quick-fix traffic campaigns such as social bookmarking to your auxiliary sites such as your squidoo pages and mini-sites. If you lose one of these sites, it is of little (if any) detriment to your main site.

Social bookmarking is relatively safe though as a basic strategy to create some initial traffic and rankings to your website, as overall it is designed as a human activity at a fundamental level. It is, however, best to be sparing with it with respect to your main website. I would recommend, if you’re looking for your site to be around for a long time to come and be able to build and maintain large volumes of traffic, that any social bookmarking campaign you run directly on your main site should be distributed over a longer period of time, as this would more accurately reflect natural human activity. For my own sites, the main website would typically trickle a bookmark in every couple of days, whereas each auxiliary page or site would build several bookmark links per day.

To summarize:

1. Create your main site and use a variety of traffic building methods. Your social bookmarking to this site should be slow and randomized over a longer period of time.
2. Create several mini-sites with useful, high quality information and make sure these sites link back to your main website. These sites should build social bookmarking campaigns to every page with a backlink to your main site and the campaigns can operate over a shorter time period, although they should still be distributed over time. Don’t blast the links out all at once or you’ll lose some of the value of the campaign.
3. Repeat this process for any auxiliary pages you build such as Squidoo, YouTube videos and Hubpages sites.

Traffic Android can help you automate the bookmarking process as it is typically quite a time consuming process to engage in.