Lessons from Yahoo News’ Failures

Chris Lehmann’s Purple Reign gives an inside view into the demise of the Yahoo News machine.  Weighing in at nearly 9 thousand words it might take you a while to get through it.  I’ve snipped out some of my favorite passages from my own perspective. The original article is a much better read than my uneven summary.

Beg, Borrow, Steal

Lehmann relates how one of his reporters acted to get his job done without overdue concern to corporate support:

My reporting team did important and groundbreaking work on … the 2010 Deepwater Horizon oil spill—our reporter moved down to New Orleans from New York, on his own dime, to cover it.

Often we have to get things done without all the support we would like–without all the resources we think we need. I like this example of dedication from this reporter who moved to the site of a big story to cover it.

An Effective Algorithm Can Be Very Seductive

I wrote recently about using metrics to manage performance. I noticed Lehmann’s story about how a successful algorythm–CORE–pushed the news into a particular direction:

“… CORE was to take the data Yahoo had collected from its registered users and develop algorithms that would accurately predict what stories would provoke site visitors to click, linger, and read…. In its first year of operation, CORE proved to be a phenomenal success, netting a 270 percent increase in pageviews.

“… front-page producers would talk of “forcing” stories that defied the counsel of the CORE algorithms—usually breaking news in the world of politics, policy, and global affairs—into placement on the front page. In other words, the independent exercise of news judgment, in this code-happy universe, was regarded as a willful lurch into manual-override mode. And even when a story was “forced,” if the click-count didn’t spike, it was quickly banished to the nether regions of the site, so the algorithms could resume their seamless work of optimized news sorting and placement.”

The passionless algorythm they had in place was very effective. So effective that not following it was very difficult.  My favorite part of the quote was this sentence:

In other words, the independent exercise of news judgment, in this code-happy universe, was regarded as a willful lurch into manual-override mode.

We must be cautious. It is difficult (impossible) to codify every value and it’s relative importance.  Without that the algorythm will always have blind spots. Human judgment can’t be fully replaced by metrics.

Honor Both Halves of Your Company Mash-up

Yahoo news tried to merge two separate disciplines. It was “Technology Company Meets News Company.” But they didn’t really understand the news business. My favorite example from Lehmann’s article is an executive that gets concerned because they were reporting a story they hadn’t already heard from other sources:

‘One of the first pieces I reported for the site concerned the tactic, favored among some militia groups, of committing crimes in order to ambush law enforcement personnel. Before the post went live, I fielded an anxious phone call from a senior manager in Santa Monica. He was alarmed that we were reporting on this practice for a simple reason: “I haven’t heard of this before.” I struggled to find a diplomatic way to explain that publishing things that readers hadn’t heard before was something that a news organization should be doing a whole lot more of: it was, in fact, the definition of “

Stay Classy

This passage just made me chuckle.

“The company’s maximum leader, a gratifyingly foul-mouthed woman named Carol Bartz, had been given the heave-ho via phone (you stay classy, Yahoo!) after the company board decided that their stock was underperforming, even by Yahoo standards. No one yet knew who would succeed her, and I had to jury-rig a set of temporary fixes as I went down my list of urgent needs, from bringing on new correspondents to cover the 2012 election to getting full-time benefits for our blog team, who had all initially come on, Walmart style, as just-shy-of-fulltime contract workers, strategically ineligible for health care. (All together now: You stay classy, Yahoo!)”

Reminds me of Steve Krug’s advice to “be a mensch” in his book Don’t Make me Think.

Your Product May Betray You

Some shops act like a developer’s job is done when her code is in production. I like the maxim “Your job is done when your code is retired.” Lehmann gives an instructive example of production systems gone amok:

“… [commenters] would occasionally descend on a completely unrelated post en masse and smear it with thousands of imbecilic racist outbursts. On one unfortunate occasion, this happened to one of our posts, and before we could reverse engineer our way out of the hate-fest, the CMS system seized up; we couldn’t shut off the comments, and we couldn’t delete the post. For weeks on end, it just sat there, oozing ugly misspelled racist obscenities….”

In a more traditional shop web developers might point at operations to fix a problem like this. In a dev-ops shop the developer has the calling and the power to fix the site until the site is retired.

Can a Big Company Act Like a Start-up?

Lehmann had an interesting definition of start-ups.

“Startups thrive on investor bets made on the promise of future profitability and, over the course of their initial runs, are case studies in capital destruction; far from ascending to Yahoo-scale market dominance, they’re expected to rapidly burn through their initial investment stakes en route to attracting more and bigger investors.”

He critiqued the idea that a part of a big company like Yahoo could call itself a startup:

“Very much by contrast, Yahoo was a lead investor in digital enterprises. It didn’t develop bold new tech properties—it merely acquired them. The company’s main claim to profitability has been the mountain of shares it was able to flip when the Chinese online shopping giant Alibaba went public. (True to form, Yahoo declined to float its enormous Alibaba winnings to fund a bold or entrepreneurial bid to revolutionize the web; instead, it spun them off into the guardianship of a dummy corporation, so as to save the company billions in tax liabilities from the Alibaba IPO.) To goad its lead executives into startup mode would be like expecting the lead actors in The Sopranos to run a marathon without notice.”

Move Fast and Break Things

Continuing his discussion of start-ups he criticized how the “move fast and break things” mantra was being applied:

“The news sites had laid off their only two fulltime copyeditors in an earlier round of staff cuts, and never bothered to replace them; why copyedit, after all, when you can break shit? That lustily intoned startup mantra, I knew, was certain to continue sanctioning colossal wastes of resources and effort.”

I had a director use that catch phrase. I saw him trying to goad an established, old company into pushing boundaries and breaking out of a rut. 

Of course, it can go wrong.

From Lehmann’s standpoint they weren’t just shaking things up or pushing boundaries. They were discarding essential quality controls and using the catch phrase for cover. Seems again like they weren’t honoring both halves of their company mash-up.

Businesses Endeavor to Make Money

The common element of a business–as opposed to a non-profit, and army, or a government agency–is to make money. News companies are no exception. News companies have developed structures to try and dampen the effect of money seeking on the integrity of the news itself–the company’s product. 

Again, in the technology plus news company mash-up the foreign sensibilities around news making weren’t respected.

“… Kathy Savitt, went into heads-must-roll conniptions when she learned that Yahoo Sports NBA blogger Kelly Dwyer made passing reference to Cleveland Cavaliers owner Dan Gilbert’s career as a predatory lending baron as CEO of Quicken Loans. Just days before, Yahoo had entered into a lucrative partnership with Gilbert’s company…. Yahoo sources told Draper that Savitt began loudly demanding that Dwyer be fired for his revenue-threatening thought crime…. This obsequious brand of synergy speaks volumes about Yahoo’s recent devolution. … Savitt had precisely zero background and training in journalism, apart from a stint thirty years before at her college newspaper. Mayer, in her infinite market-prostrate wisdom, merely added all of Yahoo’s editorial operations to Savitt’s bailiwick as marketing director—because, you see, that’s just how Yahoo rolls.”

That’s a good sum up of the dangers of running a news company when you’re really a technology company

Is It News?

The point of the article in a nutshell:

The scrupulous separation of business and reporting mandates, and the pride of journalistic craft necessary to effectively police that separation, are no doubt old-media traditions, but they became old for a reason.

Is it really “news” as our grandfathers knew it if the machinery that finds, grooms, and publishes it doesn’t have proper separation of concerns?

What Does it Mean for Me?

In our rush to create “synergy” mash-ups are we creating franken-fruit? New things called by old names but without the integrity of the old disciplines. Is the sexy technical side valued in the organization more than the discipline of the non-technical side?

This article inspires me to take care:

  • Honor both halves of your corporate mash-up, not just the sexy technology stuff.
  • Recognize when you’ve been put in charge of something you have no clue about and become a learner again.
  • Don’t hide behind start-up or agile catch-phrases like “move fast and break things.” Honor the problem. Refuse to over-simplify.
  • Don’t wait for corporate support for every step. Some-times you need to move to the gulf to cover the deepwater horizon oil spill.
  • Stay classy. Be a mensch. We’re all people.


Published by

Tyler Peterson

Tyler Peterson

Web Dev surviving a management job in the shadow of the Rocky Mountains.