Google creates its own antitrust woes with poor communication over search listings

Did Google deliberately try to reduce the rankings of ProtonMail, a tiny rival to Google’s own Gmail service? Almost certainly not. Even Proton doesn’t seem to believe that. But the case highlights how Google’s problems with publisher, business and webmaster communication can hurt it as it faces challenges on antitrust grounds.

What happened with Proton

Proton Technologies is a Swiss-based company offering a secure, encrypted email service called ProtonMail. It might be an attractive alternative for those who worry a service like Gmail isn’t private enough, either from government requests or Google’s own ad uses.

Last November, Proton noticed that they were seeing a drop in daily signups for ProtonMail. Wondering why, the company started looking into its rankings on Google and determined there was a problem. In particular, ProtonMail wasn’t showing in the top results for “secure email” or “encrypted email,” as it assumed was the case in the past.

Proton then suffered a problem that’s not unique for businesses and publishers. It had no guaranteed way to get an official answer from Google if there was a problem.

Google offers a wide-ranging toolset called Google Search Console that tells businesses if they have problems with their sites. Proton told Search Engine Land it even made use of the toolset. The problem is that the system doesn’t allow site publishers to contact Google if they suspect something is wrong on Google’s end. There’s no way to ask for help, unless you have received what’s called a “manual action,” a penalty placed on your site by a human being. Proton had no manual actions, it told us.

Without such an option, Proton ended up using Google’s spam reporting tool earlier this year. There was no indication that Proton had been spamming Google. But it appears Proton hoped that by using the form, it might trigger a review by Google which, in turn, would uncover what the real problem was.

That didn’t solve the issue. Finally, ProtonMail tweeted out for help in August to Google and to Google’s former head of web spam, Matt Cutts, who’s on leave from the company and hasn’t been involved with it for over two years. Moreover, a new head of web spam was named ages ago.

Still, reaching out to a semi-former Googler seems to have done the trick. Within about a week, the problem was resolved. Exactly what happened was never explained.

Enter the antitrust concerns

Last week, this all drew attention it hadn’t really received before because Proton did a blog post about it, one that raised the specter that it was perhaps related to competitive issues.

This incident however highlights a previously unrecognized danger that we are now calling Search Risk. The danger is that any service such as ProtonMail can easily be suppressed by either search companies, or the governments that control those search companies.

The only reason we survived to tell this story is because the majority of ProtonMail’s growth comes from word of mouth, and our community is too loud to be ignored. Many other companies won’t be so fortunate. This episode illustrates that Search Risk is serious, which is why we now agree with the European Commission that given Google’s dominant position in search, more transparency and oversight is critical.

Could that have really been the situation here?

Unlikely competitive reasons were to blame

It’s unlikely. Google has over one billion daily active Gmail users. ProtonMail has just over a million, according to its recent post. It shows no growth trajectory that’s going to cause it to rival Google even in years to come.

Given all this, would Google really have actively worked to suppress it while not bothering to do the same for real email rivals? For example, Outlook ranks in the top results on Google for a popular term like email.

It doesn’t make sense. Even Proton isn’t saying the issue was due to competitive reasons, with cofounder Andy Yen telling Search Engine Land via email:

From the data we have, it is impossible to draw a concrete conclusion. We are willing to give Google the benefit of the doubt here and in our blog post, we aren’t drawing any conclusions in this regard.

We are grateful to the individual Googlers who stepped in to fix the issue, but overall this was a very difficult and costly situation for us. We are software developers ourselves, so we know that software bugs do happen, and Google isn’t infallible either, but when Google isn’t behaving correctly, the stakes can be very high.

At the end of the day, we hope that by sharing our experience, more people will become aware of Search Risk, as it is a challenge the internet community is going to have to confront.

“Search Risk?” I’ll get back to that. But even with Proton thinking it could be a completely innocent technical glitch, this case will probably come back to haunt Google. In particular, it’s harmful as the European Union continues its antitrust review and actions with the company.

Indeed, years ago, Google almost certainly wasn’t trying to act anti-competitively against tiny UK-based shopping search engine Foundem. But spam actions against that company were the seed for other complaints and concerns to grow. Last year’s antitrust charges levied against Google by the EU grew directly out of that.

In short, Google’s can’t really afford to be making mistakes with anyone who can be deemed a competitor, because they have a big club to swing that other publishers don’t get — that of Google acting competitively. And even if Proton doesn’t swing that club, others may take its situation as an example to challenge Google.

Despite glitch, Google did still send Proton traffic

It’s a bit of a side-issue among the bigger issues here, but it’s worth addressing. Proton said it had a growth rate drop of 25 percent because of the Google change. However, it really has no idea how much the Google drop harmed it. This is because, as it turns out, Proton had no idea how much traffic Google was sending it before, after or even now.

Proton is focused on the fact that it didn’t rank for well for a period of time for the two keywords mentioned above. Unfortunately, rank checking is a terrible way to assess how well you’re doing with Google or search engines in general. Sites are typically found for many different terms. Focusing only on some is far from the full picture.

These terms might have been traffic drivers for ProtonMail or not. Proton doesn’t know directly, because the company told Search Engine Land that it’s not running any type of analytics that would show how much traffic it gets from Google or other sources.

The company said it doesn’t use Google Analytics specifically because of privacy worries. It could, of course, find this type of data without using Google Analytics, such as by processing its own server logs directly. That’s much more complicated and time-consuming, but it’s an option.

So where’s that 25-percent drop in growth come from? Proton emailed us this:

We saw a noticeable drop in the number of daily sign ups with everything else held equal. More strikingly, after Google fixed the problem, we saw a >25% increase overnight (we changed nothing on our side). For us, this 25% was the difference between bleeding money each month and being able to break even.

Keep in mind, it’s not that Google wasn’t sending Proton traffic. It’s just that the loss of ranking well for those terms, and perhaps others, caused it to get less traffic than before. That drop in traffic edged the company out of making money to breaking even. That leads to the whole “search risk” issue.

Everyone has “search risk”

The bottom line is that any business or publisher is at “search risk,” as Proton dubbed it in its post, where losing search visibility could jeopardize your business. It’s not a new risk. It’s one that literally goes back over 20 years, to the days when Yahoo was deemed the internet “gatekeeper” that could make or break businesses.

Don’t depend on search engines, Google or otherwise. For that matter, don’t build any business on the idea that you’re going to somehow get free traffic from a source, such as Facebook, Pinterest or whatever. That should be common sense. If you’re not paying for something, you’re not guaranteed to get anything.

Wise search marketers know this. Smart SEOs know you don’t want to have an overdependence on Google. Algorithms change all the time. But apparently in 2016, it’s still a lesson that people need to learn.

Google needs to improve communication

That said, Google could and should do a better job with communication. Something was wrong with the ProtonMail site, in terms of how Google was processing it. We know that, because something was fixed. Google just won’t say what. All it will say is the statement it sent us below:

Google’s algorithms rely on hundreds of unique signals or “clues” that make it possible to surface the results we think will be most relevant to users.

While we understand that situations like this may raise questions, we typically don’t comment on how specific algorithms impact specific websites. We’re continually refining these algorithms and appreciate hearing from users and webmasters.

While in many cases search ranking changes reflect algorithmic criteria working as intended, in some cases we’re able to identify unique features that lead to varied results.

We’re sorry that it took so look to connect in this case and are glad the issue is resolved. For webmasters who have questions about their own sites, our Webmaster team provides support through the Webmaster Forums and office hours.

It shouldn’t have taken so long for the problem to be fixed. Google itself shouldn’t want it to take so long. The company needs to find a better way for publishers to report potential errors and get resolutions. I wished for that back in 2006 and again in 2011, as part of my revisiting my “25 Things I Hate About Google” post:

Sure, a paid support option might put you under fire that you might be making algorithm updates like Farmer/Panda just to generate support revenue. But others might appreciate a guaranteed route.

If not paid, maybe you could give anyone who registers with Google [Search Console] one or two free guaranteed express support tickets, so that we don’t have bloggers talking about getting in contact with Google being a “crap shoot” and diminishing the huge amount of resources you do put in to support through Google Webmaster Central.

Now it’s been 15 years since I first had that wish, and it still hasn’t been solved. Yes, there would be time and cost involved. But that might be well worth it, versus adding more ammunition for those who might use glitches to attack Google on antitrust grounds.

Google’s main advice for those with problems is to use its Google Webmaster Forums. Proton was even going to try that next, it told us, if its tweets didn’t help. Personally, I’d never want someone to go there because:

1. while Googlers are there, they’re not guaranteed to review your problem; and
2. some problems (as with Proton’s) can only be diagnosed by Googlers; and
3. most people who answer in the forums are not Googlers; and
4. non-Googlers might not even give the right answer.

For instance, here’s a non-Googler telling someone they have a manual action against them because the site can’t be found in Google’s search results and the Google URL shortener doesn’t work for it. Maybe. Probably, even. But that person is guessing. The only way to actually know if there’s a manual action is by going into Google Search Console as the publisher and checking. A third-party person can’t tell you.

Still, that’s the option you have. Unless you catch someone from Google’s attention another way, as ProtonMail did. Or MetaFilter did in 2014.

[Article on Search Engine Land.]

About The Author

Danny Sullivan

Danny Sullivan is a Founding Editor of Marketing Land. He’s a widely cited authority on search engines and search marketing issues who has covered the space since 1996. Danny also serves as Chief Content Officer for Third Door Media, which publishes Marketing Land and produces the SMX: Search Marketing Expo conference series. He has a personal blog called Daggle (and keeps his disclosures page there). He can be found on Facebook, Google + and microblogs on Twitter as @dannysullivan.