Mobile Bot Traffic Reportedly Grew 30% In 2013

It’s no secret that mobile — and specifically Android — is becoming a major vector for malware. The result of this, according to a new report from CAPTCHA-based advertising firm Solve Media, is that suspicious traffic from mobile devices grew 30 percent in 2013 and currently accounts for about 25 percent of all U.S. mobile traffic.

Overall, suspicious web traffic across Solve’s network increased about 40 percent over the course of 2013 and now accounts for about 61 percent of all traffic. Just because the traffic is suspicious doesn’t mean it has to be a bot, though. Solve tells me it can confirm that about 72 percent of the suspicious traffic is definitely from bots.

That number is in line with what we heard from security CDN service Incapsula late last year.

In Q4, the highest level of bot traffic came from Southeast Asia, China and Eastern Europe, with hotbeds in Singapore, Taiwan, Poland, Lithuania and Romania.

Solve’s data is based on looking at over 700 million CAPTCHA verifications on about 8,400 sites.


Solve Media’s CEO Ari Jacoby believes that the spike in bot traffic late in the year is in large part due to the bigger ad budgets that are being spent around the holidays. “With US advertising budgets expected to top $182B by 2015, brands and publishers must both commit to adopting fraud prevention measures now,” Jacoby said in a statement today. “Aggressive botnet operators are not only stealing marketers’ hard fought budgets, they’re also creating false results on campaign performance.”

Given that Solve protects against this kind of bot traffic, it’s always worth taking the actual numbers with a grain of salt and the sites that use its system may be more prone to being attacked than a regular website, which may skew its data, too. Overall, though, the numbers in this report seem to track what we’ve heard from other companies.

Read more: Mobile Bot Traffic Reportedly Grew 30% In 2013

Mobile App Marketing Costs Hit Record Highs In 2013

As the app stores fill up, it’s getting ever more expensive for developers to acquire new users. The cost to acquire a loyal user, or one that opens an app three times, grew from $1.30 in 2012 to $1.62 in 2013. Meanwhile, app downloads over 2013 broke records, though the download growth has been slowing, as previous reports have pointed out, indicating that some maturer markets are nearing their saturation point.

However, as app marketing platform makers Fiksu point out, while the percentage slowdown may be accurate this year, it’s based on a larger and ever-increasing raw number of total downloads. In December 2013, for example, the company has been predicting somewhere between 6.6 million and 7 million daily downloads on its own Competitive Index.

The index represents the number of downloads per day of the top 200 free iPhone apps.

That ~7 million figure will be the highest volume since Apple banned automated downloads (bots) in the beginning of 2012. Those downloads are also why Q1 2012 was as large as it was – it’s likely that a good many of the earlier downloads were not real, notes the company. This could, perhaps, somewhat temper Flurry’s earlier findings, which stated that the sizable number of Christmas Day downloads seen in 2011 and 2012 (more than a twofold increase in Christmas Day downloads vs. other days in the first three weeks of December) was not visible again in 2013. But it wouldn’t be the only explanation for a cool down, of course, as the “bot ban” was enacted in February 2012 – meaning that last Christmas was already in the post-bot timeframe.

Flurry’s data could also skew more toward the long-tail, as it’s based on apps using its SDK, while Fiksu’s Competitive Index uses apps in the top 10 and plots the full curve to get to their numbers.

With more downloads overall, and a smaller spike in new device activations over Christmas 2013, a plateauing may be expected.


The App Store this year topped a million apps and smartphones became nearly ubiquitous in developed markets, making it increasingly costly for developers to acquire new users. Fiksu’s Loyal User Acquisition Cost Index, which measures the cost in acquiring new users who will open an app at least three times, has grown over the course of 2013, as well as year-over-year. In Q1 2013, user acquisition costs were around $1.40, up from 2012′s $1.25. By July, costs were breaking records, making July the most expensive month since December 2011.


By August costs grew again, spiking to $1.90, as marketers aimed to test their strategies ahead of the 2013 holiday period. By Q4, costs were at $1.84, far higher than the $1.37 in Q4 2012. Of course, the holidays are always an expensive time to market an app, but more telling is that year-over-year increase of 32 cents ($1.30 in 2012 to $1.62 in 2013).

This year saw a number of new apps in the Top Charts, including FitBit, Candy Crush, QuizUp, Temple Run 2 and others that weren’t around in 2012, each generating millions of downloads on their own. For well-known hit makers, hitting one million downloads is now par for the course.

The steady overall growth may be cause for concern – at least for smaller developers hoping to find their way to the Top Charts, which are tough for newcomers to break into. For consumers, this means in 2014, we might see more of the newer, younger companies trying darker shades of “growth hacking” as a way to find initial traction. Be warned.

View post: Mobile App Marketing Costs Hit Record Highs In 2013

Primo Is An Arduino Robot That Teaches Kids Programming Logic Through Play

Dan Shapiro’s Robot Turtles board game Kickstarter showed there is serious appetite for kids’ games that aren’t just fun to play with but also sneakily teach core coding principles. Instead of the $25,000 he was aiming for, Shapiro raised more than $630,000. Geeky moms and dads clearly have money, and will spend it on the right bit of educational kit.

With that kind of Kickstarter community response, it’s pretty likely we’re set to see a wave of educational toys doing cool fun stuff with programming principles. To wit, meet Primo: a physical programming interface that teaches children programming logic while they control the movements of an Arduino-powered robot.

All of Primo’s electronics are concealed inside wooden boxes, so from the child’s point of view they’re playing with blocks, a board and a cute little robot. But as they snap the coloured pieces (instruction blocks) into the board (the physical programming interface) they are building up a set of instructions that the wheeled bot will execute when they push the big red button. So they get to see their program come to life as the bot moves around the room and navigates around household objects.

The instruction blocks comprise four different coloured pieces: forward, to move the bot forward; left; right; and the green circular function block. The function block adds a little more complexity to the basic instruction set as it calls the last line of blocks on the board every time it’s called. Aka it’s a sub-routine.

The function element, used in conjunction with the setting of longer physical paths for the robot to complete, then requires kids to use logical thinking to build up longer sequences of instructions to complete the challenge. And that’s the subtle learning it’s hoping to achieve.

It’s certainly a lot more basic than the Kano DIY computer Kickstarter – but the idea is to offer coding ‘baby steps’, for four-to-seven-year-olds, not throw kids in at the deep end.

“Skills are mastered gradually. Mountains are climbed one step at a time. Think of Primo as the very first step in a child’s programming education. Primo provides the very basic ABC of programming logic,” Primo’s U.K.-based (Italian) creators note on their Kickstarter page.

They’re aiming to raise £35,000 to get the kit to market. The full, assembled kit costs £160 to early Kickstarter backers — or £135 for a DIY version that you can self-assemble at home. They’ve already managed to raise more than £5,500 since the campaign kicked off on Friday, with 27 days left to run. If it hits its funding target, they’re aiming to ship to backers next August.


Read the original post: Primo Is An Arduino Robot That Teaches Kids Programming Logic Through Play

Google updates reCAPTCHA to test whether you’re human before, during, and after you interact with CAPTCHAs

Google today released an update to its reCAPTCHA system that creates different classes of CAPTCHAs for different kinds of users. In short, it makes your life easier if you’re a human, and your work much harder if you’re a bot.

For those who have encountered CAPTCHAs and reCAPTCHA, but have no idea what they are, here’s a quick primer. CAPTCHA stands for “Completely Automated Public Turing test to tell Computers and Humans Apart”, and as its name implies, it is a quick test used in computing to determine whether or not the user is human. You’ve probably encountered hundreds of these if a site decides to verify whether you’re human or not.

recaptcha Google updates reCAPTCHA to test whether youre human before, during, and after you interact with CAPTCHAs

reCAPTCHA, which was acquired by Google in September 2009, is similar to the CAPTCHA interface, except that it asks users to enter words seen in distorted text images onscreen. It presents two words: one which it knows (used to test whether you are human), and one which it doesn’t (used to help digitize the text in books).

Google notes that over the last few years, advances in artificial intelligence have reduced the gap between human and machine capabilities in deciphering distorted text, and the reCAPTCHA team has been making its system more adaptive via extensive research and steady improvements.

Unsurprisingly, Google wouldn’t share too much detail as to how the new system works, aside from saying it uses advanced risk analysis techniques, actively considering the user’s entire engagement (before, during and after) with the CAPTCHA. In other words, the distorted letters are not the only test.

Here’s what Google says it gains from the changes:

This multi-faceted approach allows us to determine whether a potential user is actually a human or not, and serve our legitimate users CAPTCHAs that most of them will find easy to solve. Bots, on the other hand, will see CAPTCHAs that are considerably more difficult and designed to stop them from getting through.

Since humans find numeric CAPTCHAs (pictured above) significantly easier to solve than those containing arbitrary text, Google will be showing you more and more numbers. Bots, meanwhile, won’t even see them.

That’s not all. Google says “significant advancements” to the reCAPTCHA technology are on their way “in the next few months.” We’ll keep you posted.

Top Image Credit: Johannes Eisele/Getty Images

Link: Google updates reCAPTCHA to test whether you’re human before, during, and after you interact with CAPTCHAs