Why Chegg Is Abandoning A Business Worth Over $200 Million A Year

Yahoo folks love deals that improve margins by giving away strategic/core competencies (CEO of Chegg is ex Yahoo):

Chegg announced yesterday that it had reached a deal with Ingram Content Group, a book distributor, that will allow the company to hand over its print textbooks and dramatically improve its margins.

via Why Chegg Is Abandoning A Business Worth Over $200 Million A Year | Fast Company | Business + Innovation.

The Most Important Push Notification metric: impact on churn

The writer suggests measuring churn on your mobile app to see the impact of your push notifications.  These sort of alerts are effective but annoying for the user if the content doesn’t match the priority of the medium.

From the publisher’s point of view, some could view this simply as a way to fire their worst users so they are fine living with the uninstalls.

In fact, it is much easier for users are to just delete your app rather than to find your settings page and turn off push notifications (or to turn off notifications for your app globally from Settings.app).

Sure, sending more pushes could mean you see an increase in retention from some of your users. But can you prove that you are not losing more users by sending (more) pushes?

via The Most Important Push Notification metric — Medium.

Transition from desktop to mobile: Stocktwits edition

Continuing to collect stats on transition from desktop to mobile for various online properties:

Over the past year, the percentage of our total audience attributed to mobile (both app and mobile web) has increased from 29% to 43%. The number of messages originating from mobile has increased from 14% to 30% of all messages. More striking, of our logged-in community members, 36% are mobile app-only — a 150% increase from last year.

via The (Mobile) Investor of the Future — Medium.

The decline of blogs

Blogs are certainly in decline (or maybe moving to power law distribution) but there are many online properties (not strictly blogs) growing traffic very well on desktop and mobile (e.g. Yahoo).  To say it in another way, online and mobile traffic is growing steadily but where that traffic ends up is changing.

Inbound links from bigger sites also aren’t worth as much as they used to be, suggesting that even big sites are struggling to maintain and grow their traffic.

via Google and blogs: “Shit.” – Marco.org.

Adware alums coming out of the woodwork in this thread

I’ve worked for adware companies for a long time. I find it crazy how many news publications, investors, and technologists don’t realize how many companies are part of the adware ecosystem.

Iron Source is one of them and they just raised a lot of money from JPMorgan and Morgan Stanley to fund acquisitions in preparation for their IPO.

Adware Bundler IronSource Raises $105M to Buy Startups Ahead of IPO | Hacker News.

On how key loggers work

No malware literally logs keys typed anymore. I cannot stress that point enough. Instead they log form submissions (e.g. POST requests) which give the malware author much more useful information they can data mine in an automated way (e.g. URL, named parameters, etc). This works even on a “secure” page (e.g. HTTPS with extended certificate).

via As an aside: Everyone who read this article, please keep in mind that the proces… | Hacker News.

Opera launches App Pass zero-rating tool for carriers

The tool is a feature of the Opera Max mobile data management app and follows the release a few years back of Web Pass, which aims to help carriers provide pay-per-use generalized web access. At launch, it can be used to provide free access for set periods of time; it will soon also allow sponsored and “paid passes” to apps.

via Opera launches App Pass zero-rating tool for carriers | Gigaom.

No, you don’t need a ton of data to do deep learning

You don’t need a ton of data to do “big data”:

“While more data can be better I think this is stopping people from trying to use big data,” Howard said. He cited a recent Kaggle competition on facial key point recognition that uses 7,000 images and “the top algorithms are nearly perfectly accurate.”

via No, you don’t need a ton of data to do deep learning | Gigaom.