Why Did Google Flip-Flop On Cracking Down On “Rogue” Websites? Some Troubling Possibilities (Forbes Cross-Post)

By Eric Goldman

Earlier this month, Google announced that it may downgrade search results for a website if Google receives a high volume of “valid” takedown notices against the website.  Google’s move has confused many Google-watchers, largely because the exact implementation details are important but aren’t being disclosed.

However, I’m confused for a more fundamental reason.  Google staunchly opposed the Stop Online Piracy Act (SOPA), yet Google’s move partially implements SOPA anyway–and makes one of the analytical errors that made SOPA so objectionable.  (See my prior blog post on other recent efforts to recreate SOPA).  This post tries to figure out why Google flip-flopped on SOPA.

The Analytical Error

Theoretically, both SOPA and Google’s algorithmic change are intended to curb “rogue” websites.  Unfortunately, any definition of rogue websites creates false positives by including websites that will mature into legitimate players over time.  For example, YouTube circa 2005 might have looked like a rogue website at the time, but few folks would characterize it that way now.  Or, a site like Pinterest (with its legally questionable “sideloading”) might initially look like a rogue website until its lawyers clean things up.

In 1998, Congress enacted a safe harbor scheme to balance the interests of copyright owners and providers of user-generated content (UGC) websites.  The basic deal: copyright owners send takedown notices for specific items of user-posted infringing content, legitimate UGC websites honor those takedown notices, and the websites avoid copyright liability for their users’ content.  This scheme, while imperfect, has worked well enough to help the UGC ecosystem flourish.

SOPA threatened to undermine this balance.  Instead of sending takedown notices for individual content items, copyright owners could send cutoff notices to the website’s service providers that, if honored, would marginalize the entire website.  Thus, SOPA’s remedy didn’t fit the problem: the problem is individual infringing items, but the remedy equally affects both legitimate and illegitimate content items.

Google’s algorithmic change creates the same problem-remedy mismatch as SOPA.  Based on complaints about individual content items, Google’s algorithm may ultimately downgrade the entire website–even if all of the takedown notices are legitimate.  (Thus, I’m not even addressing the many illegitimate takedown notices competitors will send Google to try to game its algorithm).  By applying the penalty site-wide as opposed to individual content items, the search results for legitimate content items at that downgraded website also will be marginalized.  Thus, Google’s algorithm replicates one of the SOPA’s most objectionable aspects.

Why Did Google Flip-Flop on SOPA?

After fighting SOPA, why did Google choose to partially implement it voluntarily?  Some hypotheses:

1) The move helps searchers find more relevant results.  This could be plausible if, for example, rogue websites are highly correlated with other problems for searchers, such as malware, that aren’t adequately screened by Google’s 200+ other algorithmic signals. Google’s blog post didn’t tell that story, though.  Instead, Google offered only a weak explanation of how searchers might benefit from the move by seeing more “legitimate” content sources, and Google hasn’t acknowledged the countervailing risk that legitimate content may be downgraded in searchers’ results–an outcome that unquestionably hurts searchers in their quest for the most relevant results.  Indeed, Google’s algorithmic change, on a net basis, could degrade search results relevancy for searchers.  Techdirt explores this issue more.

2) Google felt this move would help reduce its legal risk.  However, Google already qualifies for Congress’ 1998 safe harbor (17 U.S.C. 512(d)), and the voluntary algorithmic change doesn’t directly improve its legal posture.

3) Google hopes the change will improve its relationship with Hollywood, which in turn could have ancillary benefits like unlocking more content deals.  Given how many folks in Hollywood believe Google is the devil, I doubt an algorithmic change will change that.

4) Google hopes the move staves off more draconian Congressional regulation.

5) Google is acting at the Obama administration’s behest to curry political favor for other Google policy initiatives.

6) Google philosophically believes rogue websites are bad and its algorithm doesn’t do enough to screen them out.

I could make arguments supporting each of these hypotheses, but I remain troubled that Google hasn’t persuaded us that its change is in searchers‘ best interests.  Usually Google’s good intent for its algorithmic changes is apparent, but earlier this year I had similar questions about Google’s motives with “Google Search Plus Your World,” which promoted Google+ in search results compared to other social media services.  Google offered a very weak explanation that the Search Plus Your World change was in searchers’ best interests, yet Google also was promoting its proprietary offering more prominently than its competitors’ offerings.  Regardless of Google’s intent, the Search Plus Your World integration created potential anti-competitive effects.

Along those lines, let me offer one last hypothesis for Google’s current move: Google flip-flopped on SOPA because the algorithmic change reduces the exposure of new disruptive marketplace entrants that compete against Google’s other UGC properties.  In effect, by (fatally?) downgrading their rankings, Google can keep websites like YouTube circa 2005 and the next nascent Pinterest from growing into bona fide competitors to Google’s franchise.  If this hypothesis is true, perhaps Google initially resisted the algorithmic change because it knew that the move didn’t benefit searchers, but Google finally acceded to the copyright owners’ requests when it dawned on Google that the move would help it suppress potential new competitors.

Conclusion

Don’t get me wrong: for years, I have vociferously supported Google’s right to use whatever ranking algorithm it thinks best serves searchers (see, e.g., my articles on search engine bias from 2006 and 2011).  However, this is the second time this year (after Google+’s integration) I’ve had to ask if Google is really trying to benefit searchers, or if it’s doing something else–such as acting like an incumbent trying to shut the door behind it.  Even if Google’s motives for the algorithmic change are, in fact, legitimate, the potential anti-competitive implications of this move are hard to overlook.