Tuesday, February 11, 2014

Facebook Releases New Data Concerning NSA



We are pleased to release new data today about the volume and type of national security requests we’ve received from the U.S. government. Along with others in the industry, since last summer we have been advocating for the right to provide additional transparency about the national security-related requests we receive. With last week’s announcement that the U.S. government has relaxed its limitations on what we are allowed to disclose, we are now permitted to provide important new information that we believe will help foster an informed public debate about the government’s efforts to keep the public safe.

The data we are releasing today should be understood against the backdrop of the transparency reports we have issued over the past nine months. Last summer, in the immediate wake of sensationalist and inaccurate media accounts of purported government access to Facebook user data, we published data showing that, in the last six months of 2012, a small fraction of one percent of Facebook user accounts were the subject of any government data requests of any kind, national security-related or otherwise. We subsequently released a transparency report for the first six months of 2013, which again showed that the total volume of user accounts that were the subject of any law enforcement requests was a small fraction of one percent. We continue to believe the information included in these reports was an important contribution to the public debate over government surveillance practices, but it was limited. We were not, for example, permitted to break down the data between conventional law enforcement requests and those related to national security, or indeed even to acknowledge that we had received certain types of national-security related requests at all.

Since then, with others in the industry, we’ve worked hard to advocate for additional transparency, and we’ve made substantial progress. Last week, the U.S. government announced that Facebook and other providers may now provide additional data on the national security-related requests received. For the first time, we can separately disclose the volume of National Security Letters (NSLs) and Foreign Intelligence Surveillance Act (FISA) requests we receive. We can also now provide a breakdown within ranges of how many of those FISA requests sought the content of accounts and how many sought non-content information (such as subscriber name). And, within each of these categories, we can also disclose within ranges the total number of users accounts specified in the requests. In each of these areas – NSLs, FISA requests, content versus non-content orders, and the number of users specified in each area – we are limited to reporting data in bands of 1000.

The chart below reflects the ranges for all national security requests effective during the reporting periods and the ranges for all accounts specified in the requests.





We plan to update this data every six months, consistent with the authorization provided by the government. These reports will reflect a six-month waiting period from the end of a reporting period, again as mandated by the government. Finally, providers that receive a "New Capability Order" are required to wait two years before including data related to those orders in their transparency reports. The government’s explanation of these requirements is set out here: http://www.justice.gov/iso/opa/resources/366201412716018407143.pdf].



The new information we are releasing today marks a significant step forward. As we have said before, we believe that while governments have an important responsibility to keep people safe, it is possible to do so while also being transparent. We will continue to advocate for reform of government surveillance practices around the world, and for greater transparency about the degree to which governments seek access to data in connection with their efforts to keep people safe.

Update On Platform Updates For Facebook For April 9th

Here are the upcoming changes for Facebook due April 9th.

Table of Contents

Please note that in the below, deprecation means the field will no longer be supported by may still exist. Sunset means the field will officially be removed and no longer supported.

Targeting

The overall concept of targeting has been redefined as having four key areas (locations, demographics, interests, behaviors). This organization creates more flexibility and precision in audience creation in terms of the and/or logic between groups.
Example: historically, if you selected "Parents" "Photography" "Photo uploader" within interests, the audience constructed was People who are Parents OR interested in Photography OR people who upload [many] Photos. However, the intended audience was: "Parents, who are interested in photography and upload pictures". Now, the specific audience will be targeted as intended.
During the period allotted for the breaking change, all advertiser accounts managed by PMDs will NOT be added to the new targeting types, unified interests (unless previously added by request), or demographic and behavioral changes. This is to prevent disparity between Facebook UI and PMD tools. Access to the new features will be turned on once the respective PMD has confirmed they have altered their API to support the changes bysubmitting a request.
If you have new ad accounts creating/editing ads in both Facebook interfaces and your tool between now and when the API integration is completed at your end, please submit a request to have the user IDs of the ad accounts restricted.
As you are done integrating the new API specs, let us know so we can add your clients to the updated Facebook targeting offering.

Unified Interests

Interest targeting has been redefined to create a unified definition for what indicates interest in a particular topic. Legacy targeting concepts such as Precise Interests (Keywords, #Topics) and Broad Categories have been unified and as a result have only one targeting segement for each term (no duplicate "Baseball" and "#Baseball"). The result is one, clear definition for each targeting segment.
If you are a member of the PMD News group, please refer to this existing post for more details

Geographic targeting

  • The new Geo-Targeting API adds targeting flexibility and incorporates inclusion/exclusion logic.
  • Select any combination of Country, Region, City or Zip without restrictions.
  • Exclude locations from your targeting parameters.
  • All previous capabilities, such as radius targeting on cities, continue to exist.
  • Geo-targeting logic is now "OR" and not "AND". For example, to target California, the previous logic would have required you to specify "US" as the country and "California" as the state. With the current logic, if you specify "US" as the country and "California" as the state, the targeting will result in the entire "US". To target "California", you now only need to specify "California".
  • With the new API, the keys have changed for regions and zip code targeting.
  • IMPORTANT NOTE: The Targeting Spec for any ad created with the new geo-targeting logic (e.g. via Create Flow or Power Editor) will not be editable via the current targeting API. All other parameters will remain editable.
If you are a member of the PMD News group, please refer to this existing posts: original andupdate

Advanced Demographic and Behaviors

Advanced Demographic targeting options will now be available. Key features include better coverage of workplace, education and job title types as well as expanded Relationship status types. There is also the added functionality of selecting the recency of change of a life event (3 months, 6 months, 1 year).
A new targeting type, Behaviors, has been added. These categories are specific to a user's particular actions or past purchase behavior, past purchase behavior, purchase propensity. This category is made up of some targeting segments previously categorized as Interests.
Facebook will begin alpha testing, in the US, of advanced Demographic and some Behavioral targets in early February. The new UI with these features will not be widely available until we have finished General Availability release in mid/late Q1.
See Advanced Demographics and Behaviors for documentation.
During the month of January, PMDs will only have access to the Advanced Demographics & Behaviors APIs via whitelisted test accounts for the purpose of building and testing. Submit a request to whitelist any test accounts. To ensure parity for advertisers across interfaces, the migration will be only become available in early February when Facebook begins testing, in the US, our new UIs.

Creative Data Model

In an effort to better align our Ads API with our Ads Simplification effort, we're making changes to simplify the creative data model to be more intuitive. The focus of creatives will be on the promoted object and the creative assets instead of the creative type, making them more consistent across promoted object types.
As part of a simplification of the ad creative object, we will be removing the field type and standardizing on field names. We will infer what type of ad you wish to create based on the other fields within the creative spec.
Some notable changes:
  • type will be sunset
  • The auto_update field will be sunset. Existing ads will continue to delivery, but will no longer update after 4/9.
  • See the ad creative documentation for exact mapping field name changes.
  • For type 12 premium ads, we will infer using the locations field. Premium standard ads will not look different than standard ads anymore (previously they were 110x80px vs. 100x72px for standard ads).
  • App ads will no longer require an object_id field, so you must explicitly specify a tracking and conversion spec.
  • Facebook will sunset the creation of sponsored stories
    • Page post and page like ads already automatically have the best social context (likes and comments) added. You may choose to add share social context by specifying social_prefs=['allow_shares'] on the adgroup. Existing page post and page like sponsored stories will continue to deliver so you must support fetching them.
    • Domain and open graph sponsored stories will no longer be allowed to be created. Existing domain and open graph sponsored stories will cease to have delivery after April 9th.
  • Facebook will sunset the creation of type 4 ads. Instead, create a canvas app install ad (type 32).
    • Please note that the image dimensions of a type 4 ad may differ from a type 32 – type 32 ads must be 1200x627. It’s recommended to resize the image if you need to reuse it.
    • image_hash is a required field of type 32 ads, whereas they were optional for type 4
    • Existing type 4 ads will continue running until they are naturally phased out, so you must still support fetching them.
    • You will not be allowed to update fields of an existing type 4 creative
    • Please see below for the mapping of type 4 fields to type 32 fields.
type 4 fieldtype 32 field
object_idobject_id
bodybody
namename
titletitle
link_urllink_url
image_file or image_hashimage_hash
url_tags

Objectives

Adgroups will have an optional objectives field. While this field is optional to specify, editing an ad that already contains this field will fail if the adgroup no longer passes thevalidation requirements.

Ad Image Cropping

Facebook has added the option to specify image crops through the Ads API. Image crops are used to describe the way that an image should be displayed in each aspect ratio of the different ad placements. During rendering, the image will be cropped according to the specifications given and if no specification is provided for a certain aspect ratio, the image will be displayed in the default way.
This is optional but if you choose to specify an image crop, you should provide information for all possible aspect ratios in our platform. For more information, please check the docs.
You should also support passing through image crops for ads created outside of your tool. After the migration, if you do not pass through the image crop value, any image crops will be removed.

Enum int->string changes

This migration changes the response of certain fields from an int value to a string value. This applies to the following fields:
As part of this change, we will also be removing the subtype_name field of custom audiences

Sunset disapprove_reason_descriptions

The field disapproval_reason_descriptions is being renamedadgroup_review_feedback. The format will be changed from an array of string to an object key-value-pair of reason enumeration and descriptive text. The previous response would be :
"disapprove_reason_descriptions": [
            "Descriptive text"
         ]
The new response will be :
"adgroup_review_feedback": {
            "REASON_ENUM": "Descriptive text"
         }

LinkedIn Acquires Bright


Parker Barille talks about the acquisition of Bright.

We recognized instantly that LinkedIn and Bright have a lot in common. For example, we’re both passionate about connecting talent with opportunity at massive scale and obsessed with increasing the effectiveness of the job seeking and recruiting processes.

At LinkedIn, we’ve invested a lot of time and energy into developing products like “Jobs You May Be Interested In” and LinkedIn Recruiter that use data to match members with the right opportunities and employers with the right prospects at scale. And that’s why Bright is such a great fit. Its incredibly talented team takes the same data-driven approach to connecting people and employers, and has built a powerful matching technology that will help accelerate our efforts on multiple fronts.

Ultimately though, our vision is to create economic opportunity for every member of the global workforce — all three billion of them. Doing so will require increasing the volume of job opportunities available on LinkedIn. As we add more job listings over the next several years, Bright’s powerful matching technology will be integral to ensuring that the prospects we suggest to employers and opportunities we surface for prospects are increasingly relevant.

Monday, February 10, 2014

LinkedIn Removes Intro



Aimee Chanthadavong reports that LinkedIn continues to fine tune its services, and has announced on its blog that it will shut down three services it currently provides.

The professional social networking company said it aims to concentrate on doing "fewer things, but better".

As of March 7, 2014, LinkedIn Intro, which was initially launched last year to bring LinkedIn to the email inbox of iPhone users, will no longer be available. Users of Intro will be able to uninstall it between now and March 7, and switch back to their previous mail accounts. Members with Gmail accounts can continue to use Rapportive.

The decision to shut down Intro, a proxy service for emails that injects LinkedIn information for contacts, follows controversy upon the launch of the service that it was breaching privacy rights, such as making changes to users' security profiles on their devices. However, the company's senior manager for information security, Cory Scott, at the time, defended the service and said Intro was rather a potential remedy to the standard inbox.

Slidecast, a service that enables SlideShare members to upload presentations with audio, will also be shutting down on April 30, 2014. Slidecast users will be able to download their Slidecasts until then, and can continue to share the non-audio portions of their presentations with their network on SlideShare.net.

The company also plans to eliminate the support for the LinkedIn iPad app on iOS versions older than 6.0 as of February 18, 2014, to encourage members to have the latest version of its app. This decision follows LinkedIn's decision to stop supporting the LinkedIn mobile app for iOS versions older than 6.0 and Android versions older than 3.0 late last year.

Despite these decisions, LinkedIn reinforced that it is committed to looking at new ways to improve its members' existing experiences. Back in November, the company introduced Showcase Pages, a new scheme for major brands that breaks down their product portfolios and puts the spotlight on more brands.

LinkedIn also recently recorded that government requests of members' data has doubled.

Google Passes Exxon to Be No. 2 US Company



ABC News reports that Google has passed Exxon to become the second most valuable U.S. company by market capitalization.

According to FactSet data, the Internet company's market capitalization surpassed that of oil company Exxon Mobil Corp. last week. As of Friday's market close, it sat at $395.42 billion compared with the oil company's $392.66 billion.

Shares of Google Inc. have been on a steady climb since the beginning of 2013, gaining 66 percent.

Market capitalization is the number of outstanding shares multiplied by their value.

Meanwhile, Exxon's have risen just 5 percent. Since the beginning of this year, they've lost about 10 percent of their value.

Both companies trail Apple Inc.'s market capitalization of $463.55 billion.

Friday, February 7, 2014

LinkedIn Plunged 11%

CNN reports that shares of the professional networking site plunged 11% in after-hours trading Thursday after the company posted solid fourth-quarter earnings and sales but offered a disappointing outlook.

LinkedIn (LNKD) said it expects sales to rise only slightly in the current quarter, to between $455 million and $460 million. Analysts surveyed by Thomson Reuters had expected sales of $470 million this quarter.

The rest of 2014 won't be too much stronger, with LinkedIn predicting total sales of between $2.02 billion and $2.05 billion for 2014. That's far below analysts' forecast of $2.2 billion.

LinkedIn's sales are also growing at a slower clip than its competitors. The company reported fourth-quarter sales of $447 million, up just 13.8% versus the quarter prior.

Facebook (FB, Fortune 500), by contrast, posted strong fourth-quarter sales of $2.6 billion, driven by its success on mobile devices. Facebook's revenue rose 28% versus the third quarter. Twitter (TWTR) brought in a solid $243 million in fourth-quarter sales, up 44% versus the third quarter.

But LinkedIn's membership is growing faster than its rivals. In the fourth quarter, LinkedIn said membership rose to 277 million, up 6.9% from the prior quarter.

Investors punished Twitter earlier this week following its first quarterly report as a public company, after its monthly user base grew just 3.9% quarter-over-quarter to 241 million. Facebook also reported tepid fourth-quarter user growth, with its monthly user base rising 3.4% versus the quarter prior to 1.2 billion.

Of course, the comparisons between the three companies aren't exact.

LinkedIn relies less on advertising than Twitter and Facebook, for example, and earns a significant portion of its revenue from premium subscriptions -- 20% in the fourth quarter, to be exact. That equated to $88.1 million, up 10% versus the three months prior.

The company also relies heavily on its "talent solutions" business, which provides corporate recruiters with tools to promote jobs and search for candidates. That unit generated $245.6 million in fourth-quarter sales, more than half of LinkedIn's overall revenue and a gain of 9% versus the quarter prior.

Twitter Stock Tanks



Dan Nakaso and Jeremy C. Owens says Twitter's high-flying stock came crashing back to earth Thursday, plummeting more than 24 percent after the social network's first earnings report as a public company revealed lingering problems trying to expand its user base.

Some industry analysts, such as Arvind Bhatia at Sterne Agee, downgraded Twitter to "underperform" Thursday after earnings results that "are likely to raise questions on how mainstream the Twitter platform can be in the long term." .

In Twitter's first quarterly earnings report since its November 2013 initial public offering, analysts expected the company to report that it had 249 million monthly users. Instead, Twitter had an average of 241 million monthly users -- a 5 percent increase from the third quarter.



Karsten Weide, vice president of media entertainment for IDC, called Twitter's sluggish user views "very worrying."

"Twitter's stock was going down because even though revenue is growing very nicely, its user numbers are not," Weide said. "The concern is that users may not seem to think that there is enough utility to adopt the service. And the market is catching up."

Twitter stock closed Thursday at $50.03, down 24.2 percent, and hovered below $50 in after-hours trading -- a long way from its Dec. 26 peak of $73.31.

Brian Blau at Gartner said, "Clearly investors had overly ambitious expectations for Twitter. Maybe they were overly ambitious in reading the tea leaves and didn't read them right."

Financially, Twitter beat analysts' revenue expectations, with a net loss of $511.47 million, or $1.41 a share, on revenue of $242.68 million. That represented a revenue gain of 116 percent from the same quarter a year ago. After adjusting for one-time charges, Twitter reported profit of $9.77 million, or 2 cents a share.

Also on Thursday. Mountain View's LinkedIn reported a solid financial performance that beat Wall Street expectations. But the online professional networking service indicated its growth is also slowing, worrying investors.

As a result, shares of LinkedIn were down more than 8 percent in after-hours trading Thursday.

During Twitter's conference call Wednesday, CEO Dick Costolo assured analysts that Twitter will work to make the service easier to understand for new users to try to expand its user base.

"We have a very, very clear road map across a number of dimensions that we will use to drive interaction and engagement and make it easier for a broader audience to get Twitter," Costolo said.

"Twitter's value is in the future profit," said Roger L. Kay, founder and president of Endpoint Technologies Associates. "But as it gets overheated, at some point some people might say it's not being supported by anything that looks like real earnings."