AOL analytics

AOL logo - old1 In the late 1990s, “America Online” was the shiny new company everyone watched, feared and tried to copy.  Just “AOL” now, it’s hardly as fresh or inspiring. With its new CEO, logos and use of web analytics to select the stories it covers and evaluating its reporters, has AOL once again become a news organization to watch?

AOL logo - new AOL’s announcement that it will employ “judicious use of Web-analytics software” sparked the expected flutter of coverage.  It’s admitted to using data to inform (dictate?) news decisions, so you could be led to believe that AOL is adopting a true audience-based approach.  However, after reading the Feb. 22 story in BusinessWeek and the reactions gathered by Media Post News, it seems like AOL is still using a traditional advertising-based mass media strategy.  It’s still trying to be all things to all people.  It’s just using web analytics to decide what those things are.

“Audience growth and audience engagement have to be the things that we judge the most off of our journalist investments,” AOL CEO Tim Armstrong is quoted as saying.  So far, so good.

AOL logo - swirl Armstrong also said that “brand ads should be a lot bigger on the Internet today,” talking about how online advertising revenue should pick up.  But there was no mention about AOL’s own brand strategy, something that would answer the question of “What is AOL?” for audiences and advertisers once and for all.  On which niches will it focus?  How much of its content will be unique and compelling enough to those niche audiences so that they’ll come back regularly?

Patricia Handschiegel, who blogs as Daily Patricia, sums it nicely:

AOL logo - tongue “The right approach to the content business is to KNOW YOUR AUDIENCE, or
the people that come to your site, and create a product for THEM. AOL’s
approach is clearly not centered on this….it’ll drive up page views and therefore, revenue but that’s not
likely to last as the industry becomes more analytics savvy. Today, a
million uniques with zero session times, high bounce rate and no repeat
visitors isn’t seen as a sign of a lack of audience but in the not too
distant future it will.”

I haven’t researched AOL myself, so I don’t know if all of the details in the BusinessWeek and Media Post News stories really reflect what AOL is doing.  So I’ll just note some some things news orgs should think about when using web analytics to inform news decisions and evaluate journalists.

  • Evaluating success (either a site’s or a journalist’s) by total page views doesn’t work. A large number of page views may just indicate visitors got there by mistake or clicked around trying to find something.   Plus, dynamic content (Flash, etc.) will not be counted as page views. Page views can be a useful metric, but only when combined with other metrics – such as ratios – that give context.
  • Engagement can’t be determined by web traffic or behavioral data alone.  Attitudinal research is essential to find out why or why don’t people come to a site regularly, what they want and what they’re not finding.
  • If journalists are going to be held accountable for web traffic and audience engagement will they also have control over the factors that drive traffic, such as design, navigation and marketing? Or will they just submit their stories and hope for the best?
  • Money “AOL is even considering sharing a portion of quarterly profits with staffers whose work fetches the most page views.”  BusinessWeek

How will traffic goals be set?  If journalists will be rewarded for generating “traffic” (however it’s defined), will they be fired if they don’t? Will the benchmarks or starting points – and the time journalists have to reach the required traffic levels – be based on whether a topic is already established or whether it’s one a news org wants to nurture and grow because the topic is essential to achieving its strategy?

  • “Tacked to the newsroom walls in AOL’s downtown Manhattan headquarters are pages and pages of Web traffic data.” BusinessWeek Stacks.bill.pages.giUh, this would cause even me to shut down.  It’s definitely not “judicious use of Web-analytics software.”  Does AOL have a few key performance indicators that everyone understands and on which they can focus as a team?

    Software and reports don’t make decisions; people do. Successful use of web analytics depends on the decision-makers understanding and using the information correctly.  If news orgs believe the success of their websites depends on being truly audience-focused then they must also ensure the analytical resources and processes are there as well.

AOL logo - fish AOL may stumble again but at least it’s trying something different.  I look forward to learning from AOL whether it succeeds or fails.

The audience determines the value of content, not publishers

Why do most newspaper publishers persist in thinking audiences will come back to print?  Why do publishers put their own opinions ahead of their audiences?

You can’t get people to pay for anything if

  • they don’t need or want it
  • they can get it from somewhere else

You really don’t need a study to know that publishers value their content more than audiences do, but I’m glad the American Press Institute did one anyway.

API’s study, as reported by eMarketer, shows that news providers “were more likely to say that their content was ‘very valuable,’ while readers tended to settle on ‘somewhat valuable.'”

An opinion of “somewhat valuable” should not lead to a decision to charge for content, especially if 52 percent of newspaper site readers said that finding a replacement for newspaper site content would be “very easy” or “somewhat easy.”

What’s more discouraging is that 75 percent of publishers believe readers would go back to print.

What?  Why would publishers think this?

Because many of them are still looking at online as a necessary evil (my words, not API’s, obviously).  Twenty-eight percent of publishers think that they should charge for content to preserve print circulation.  Four percent want to do it to replace “lost display ad revenues.”

Eighteen percent think charging for content will establish “value for copyrighted content.”  I hope this means that before they set prices they will ruthlessly assess which content topics do indeed have value – to the target audiences.  I hope they are equally ruthless in identifying their competitors’ strengths and weaknesses, both current and potential.

Prices shouldn’t be based on

  • what the publishers think the content’s worth
  • what would make online pay for itself
  • what’s needed to make up for lost print revenue

Setting prices and revenue goals based mainly on internal reasons is like trying to get back the purchase price and remodeling costs of a house that you love but which few others do.  With anything involving money, the market sets the price.

 

 

Short-form videos need short ads; video subscriptions

As you add more short-form videos to your site, watch the length of the ads in proportion to the length of the videos.

At the Online News Association conference last week, many presenters stressed the importance of posting more short videos (30-120 seconds) more often rather than waiting days or weeks to craft a traditional long-form TV story package.   Chet Rhodes, the deputy multimedia editor for breaking news at washingtonpost.com, said that the Post has problems supplying enough video inventory to its advertisers.

Many studies, including this new report from eMarketer, show that “online video viewership has never been higher.”   However, the study points out that audiences’ acceptance of video advertising is dependent on the “growth of professional content” and targeted, less intrusive ads.

Has the growth in your video traffic – as measured by multiple metrics including the number of viewers and how much of the video was viewed – kept up with your audiences’ hunger for more live and breaking news videos?

Continue reading “Short-form videos need short ads; video subscriptions”

Total unique visitors and paid content

Because it’s easy to gather and it looks like circulation and readership, the number of monthly unique visitors continues to be a key indicator of online success for news orgs.  This is really dangerous, especially if used to develop news business models.

The total number of monthly UVs just doesn’t give any information about how engaged audiences are.  Let’s say you have 100 million monthly uniques, as paidContent.org reports the new Steve Brill Journalism Online venture is aiming for.

This number doesn’t tell you whether those 100 million of those visitors visited once or 10 times, or whether they went to one page or to 20.

You really need to know the level of engagement to sell online advertising.  And, you really need to know how engaged people are if your business model depends on paid subscribers or content.

According to paidContent.org, Journalism Online is counting on about 10 percent of its news affiliates’ audiences to pay for content.  Sounds like a realistic, reasonable number, right?

No, it’s faulty business logic.  Simply assuming a small percent of any total audience will do anything is really dangerous, and something that savvy entrepreneurs know or learn in Marketing 101.  “There are 100 million people living in this area of the U.S.  If I build a better mousetrap that costs $1, and if only 1 percent of those 100 million buy my mousetrap, I’ll have a million dollars!”

First, not all 100 million care about trapping mice.  Others won’t pay even $1 for it.  Still others don’t live near a store where they would be sold, and wouldn’t order it online or by other ways.

Estimating audiences is an art and a science.  Estimating the audiences for paid content involves more art than science, but I hope news orgs will start with understanding what online audiences want.  It doesn’t do much good to set these types of numbers based on what the news orgs need to desperately meet their revenue goals.

Using stats from Facebook’s self-serve advertising program

Even Facebook doesn't have demographics and user info. on all of its users, as this ReadWriteWeb story illustrates.  In particular, look at the comments which point out all of the problems with using an incomplete set of data.

Facebook doesn't require users to provide age, geographic location and other basic demographic data when they sign up.  Thus, the user data you get when you go through the self-serve advertising program doesn't include all of those people who didn't submit data.  And you don't know how many, or how these missing people differ.

Also, Facebook doesn't require people to update their data, which causes other problems.  One of the ReadWriteWeb story commenters noted that  students who graduate from a high school or college may or may not be counted as students.

Facebook data should not be used for planning your social media strategies and services.  You're better off guessing than using this data.  Bad data may

Continue reading “Using stats from Facebook’s self-serve advertising program”