x

Surviving the (Not Provided) Apocalypse

Not having keyword referrer data is not the end of the world… we promise! Learn what changes you can make to your reporting to account for this new reality.

The era of 0% organic keyword data is upon us.

While SEOs have had nearly two years to prepare for the inevitable referrer data apocalypse, Christmas came early – and everyone got a lump of coal. Google has made the move to encrypt search – and subsequently eliminate keyword referrer data – for nearly 100% of all organic searches.

notprovided

How did it all happen?

(Not Provided): An Abbreviated Timeline

May 21, 2010 – Google launches encrypted web search, stripping out referrer strings. Usage is very low, and the community at large barely takes notice.

October 18, 2011: Google announces that users who are surfing securely (behind SSL) or who are logged-in to Google properties (Gmail, Google+, Google Calendar, YouTube, etc.) would not have their “referrer” keyword data reported. Matt Cutts & the Google gang claimed the impact would only be single-digit.

January 2012: “Search plus Your World” entices more users to stay signed-in to Google properties.

March 2012: Firefox announces that all users using Firefox 14 and higher will now switch to HTTPS Google search by default, adding to the loss of data. At that time, roughly 14% of searchers used Firefox 14+ worldwide.

October 18, 2012: Study by Conductor finds that Google searchers are now 3x more likely to be signed in than Bing searchers.

January 2013: Google announces Chrome 25 will be SSL encrypted, ensuring further loss of data. About 35% worldwide use Chrome, so this is a significant loss.

September 23, 2013: The data apocalypse reaches a fever pitch as Google works towards making ALL searches secure – except, of course, their beloved advertising clicks. Google confirms what we’ve all suspected: Single-digit (not provided) was an outright lie.

“We added SSL encryption for our signed-in search users in 2011, as well as searches from the Chrome omnibox earlier this year. We’re now working to bring this extra protection to more users who are not signed in.”

Now that we know 100% (not provided) numbers are imminent, how can you weather the storm, talk to your teams and change your reporting?

Three Responses to (Not Provided) That Aren’t Helpful

1. Complaining.

Go ahead, get it out of your system. Close your office door, curse, headbutt a monitor. But beating on the hood won’t un-crash the car.

Your team might sympathize and share in your complaining, but that isn’t leadership. Your management will echo and intensify your concern, but they won’t respect your resolve. Complaining will only make this worse.

2. Praying for Change

Remember that time a major search giant reversed a public move they made under the guise of “privacy”? No, me neither. And if you think the public cares enough about this to suddenly surge to Bing, I’ve got a cheese farm on the moon to sell you. You can cross your fingers, sign petitions, tweet angrily at Matt Cutts… but this is your reality now.

3. Ignoring the Problem

Burying your head in the sand is a sure-fire way to get fired, replaced, or lose the trust of your team.This (not provided) problem was foreseeable. If you’ve been procrastinating on creating a contingency plan or banking on having more time to adapt to the changes, it’s time to wake up and smell the coffee.

SEO ain’t dead, but it’s limping a little bit. It’s time to talk about survival – what happens next?

Reporting Methods & Benchmarks That No Longer Work

For most SEOs, an important measurement of success was traffic generated through unbranded keywords. Keywords containing brand were typically filtered out, as companies tended to rank for their own brand names regardless of SEO.

Some of the metrics that now need to be thrown out include:

  • Traffic-per-keyword
  • Growth across keyword verticals. For example, businesses with many locations may have tracked unbranded phrases that contained specific localizers, like “Los Angeles Car Insurance” or “Texas Home Insurance.”
  • Converting keywords or revenue per keyword. You can no longer see which keywords or keyword verticals drove the highest conversions.
  • Keyword opportunities. If you’ve been using your analytics to identify new areas of content to invest time in, you’re out of luck (though I’ll show you how to go about this in the new (not provided) world in a moment).
  • Unique, unbranded phrases. Many SEO’s tracked the growth in the total number of unbranded phrases that referred traffic as a means of understanding total reach.
  • Any keyword-specific data. Not bounce rates, page views, visits, time on site… nothing.

To be clear: There is no longer a clean, succinct way of capturing branded and unbranded keyword data – and actually, that’s okay.

Keyword Data: Not All it Was Cracked Up to Be

In a presentation that should be mandatory reading for everyone who uses analytics, Annie Cushing’s  “Breaking Up With Your Keyword Data” breaks down why data broken down by keyword or non-branded organic traffic is “junk data.”

Her argument is a good one: When you start carving up data at the exclusion of things like Chrome 25, Firefox, the Google Local carousel or logged-in Google users, you get some sort of Frankenstein that is NOT a representative sample of your data.

If a data sample isn’t representative, you cannot use it to chart trends or understand what’s really happening.

In other words, the year-over-year unbranded keyword reports you’ve been using are bad data, and shouldn’t have been your metrics to begin with. Not provided has long made that data unreliable estimates at best.

Of course, knowing that won’t ease the sting, so let’s look at something a little more helpful:

What can be Measured Now?

Because we can no longer use unbranded keyword data to gauge our SEO success, it’s time to think about new metrics and mash-ups. Much of this is covered in Rob Ousbey’s brilliant “Keyword Analysis in a World of 100% Not Provided”, a useful post well ahead of its time.

The new key metrics you’ll want to watch are:

  • Amount of organic traffic coming in on a page-level
  • Average rankings across a broad array of phrases
  • The landing pages corresponding to organic rankings
  • Conversions and conversion rates of key internal landing pages

Let’s dive into the “how” and “why” behind some of these metrics.

(Not Provided) + Landing Pages = Insights

While we can no longer see which unbranded phrases brought people in, we CAN mash-up information to get an idea. Though we can’t know which unbranded phrases brought people in, we can begin to infer this information by looking at the landing pages that rank for unbranded keywords.

In this piece on Econsultancy, Dan Barker walks you through a filter you can apply to Google Analytics to automatically pair instances of (not provided) with their corresponding landing pages and report this within the organic keywords report. Once applied, it looks like this:

hack-after

Instead of seeing a single (not provided) metric, you now have a means of gauging where that traffic arrived, which you can then turn around to estimate the content of that query. From this, you can begin to infer how well your landing pages are pulling in traffic from your targeted keyword vertical.

It’s not an exact science, but it’s a reasonable indicator.

EstimatING Branded Traffic

If you think you really need to know how branded traffic is performing, you can infer it from other sources. In a tip from Rand Fishkin on his emergency Whiteboard Tuesday, there are a few ways to estimate what percentage of (not provided) might be branded:

1. Bid on broad-match brand terms with Google AdWords to calculate the approximate number of impressions and deduct from total traffic (though this can quickly get expensive)

2. Use Google Insights/Trends (which is a bit loosey-goosey)

For a very approximate understanding of branded vs. unbranded traffic, subtract these numbers from your total. Keep in mind that while this may give you an approximate trend, this is information is likely not representative of your total (I can feel Annie Cushing cringing from here!)

Understanding WHAT IS (AND ISN’T) DRIVING TRAFFIC

There are a few methods at your disposal:

1. Use Webmaster Tools Keyword Reports
While in the past the rounding and averaging of Webmaster Tools (WMT) made it a poor resource, the data has improved (somewhat) with time. Export WMT data regularly (it only lasts 90 days) to see

  • Impressions from various keywords
  • The corresponding ranked pages of those keywords
  • Estimated search traffic from those keywords
  • Average ranking for those keywords across time

wmt

Once again, you can use this data to get a feel for how well your SEO is performing.

2. Track a Broader Array of Phrases (and Their Landing Pages!)

It feels sort of ironic to turn to rankings – a metric the industry has fiercely debated the validity of – to try and aid us in our measurement plights. Still, for the time being (see the final paragraph of this guide to understand why this won’t last forever) we can cross-reference traffic volumes and ranking information to get an idea for how your site is performing.

In the past, you might have gotten used to tracking “head phrases,” the high-volume, more obvious keyword phrases.

Now, it’s helpful to track a broader array of phrases, including head phrases, mid-range phrases and long-tail phrases to help gauge how your site performs on the whole.

3. Use Tools like SEMRush

SEMRush can be a helpful guide in determining which phrases are driving impressions and which pages rank for those phrases. You can use tools like this in conjunction with other data (rankings, landing page traffic, conversions) to get a sense for what might be driving traffic and what areas you could stand to improve in.

semrush

4. Use the landing page mash-up described earlier.

Identifying New Content Opportunities

One of the key uses of keyword data was identifying new opportunities for content creation. Some of the ways to go about this now include:

  • Tools like Ubersuggest. These tools list several different queries users make when hunting for information. Use the data you get from ubersuggest in correlation with ranking and keyword volume data to identify new areas of focus.
    ubersuggesat
  • As mentioned earlier, Webmaster Tools will give you some idea of the keywords your site receives impressions for and the average ranking of your site across time. You can use this information to find areas of information where you could use an improvement.
  • I hate recommending feeding the beast, but you could use Adword campaigns to test new markets and gauge interest through clicks and impressions.

Talking About (Not Provided) With Leadership & Teams

Whether you’re in-house, leading a team or reporting to a client, there are people that this change will impact outside of yourself. How can you succinctly explain what’s happened and why it matters?

Give it to them straight – and in plain English.

Don’t start jibber-jabbering about referral strings and encrypted search – leadership may not speak that language. To understand the (not provided) changes, your teams will want to know just two things: What happened and why.

The fastest explanation I’ve heard that covers these without leaving a lot of questions is as follows:

“In what they say is a privacy measure, Google no longer allows businesses to see which keywords brought visitors into their websites, unless those visitors clicked a paid ad.  This means changes to how we report and measure success.”

Only delve into the history or tech behind (not provided) if asked about it. While it’s highly interesting to you, it may not be to them. Once you’ve established what the changes were, it’s time to talk about impact.

Explain that the problem is universal – and permanent.

Leadership needs to understand that (not provided) is not a penalty or detriment that’s unique to your business.

Nobody likes being the bearer of bad news, but it’s important that management understand that (not provided) is a Google-driven initiative, and not a result of you doing your job poorly.

  • Clearly explain that (not provided) is here to stay.
  • Articulate that this is not a penalty or a result of unethical business practices.
  • Communicate that every business has been impacted this, and that anyone who once reported on keyword-specific metrics is now on the hunt for new ways to capture impact.

Demonstrate how 100% (not provided) will impact your Processes and reporting.

Once they understand the gist of the problem, leadership and teammates care about how this will impact their jobs and the business as a whole.

Before heading into the conversation, determine the real impact this will have within those contexts.

  • Which keyword-based metrics you had been reporting on in the past, and how those metrics were used (both internally and externally)
  • Which reporting methods will need to be changed or replaced?
  • What gaps will be left in the data that simply cannot be closed, and what does that mean for the way success is measured or benchmarked?

Present a plan for the future.

Don’t just come with a problem – offer up solutions. You may not be able to get the same data or measure success the same way, but that means evolving, not curling up in the fetal position.

  • What new (or old) metrics will you measure?
  • What are the advantages and limitations of those metrics?
  • How will those metrics be measured? Have a process in hand that can be quickly explained.
  • How will those metrics be reported – and to whom?

Use this as a chance to push for greater collaboration.

Turn the problem into an opportunity to work more closely with other departments and to expand leadership’s perspective on what SEO is and does.

This is particularly important if you’re in an environment that still equates SEO to “rankings and keywords.”

Now is the time to have discussions about tying traffic into conversions; about working closely with marketing on more holistic campaigns and about tying SEO into other initiatives instead of treating it like a hands-off, tech-only discipline.

This could be your chance to:

  • Start vouching for content marketing as a means of capturing, nurturing and converting leads
  • Throw more weight behind defining and segmenting personas, looking for the “why” behind visitors to your website instead of just what brought them in
  • Turn attention to improving the overall conversion rate optimization of the website
  • Discuss closer relationships with PPC and paid advertising channels to share data and learn together
  • Reposition SEO as a consideration that needs to be made across all marketing efforts, not just “rankings and keywords”

Planning For the Future

If (not provided) really ruined your day, the following quote should scare you a bit:

“My vision when we started Google 15 years ago was that eventually you wouldn’t have to have a search query at all. You’d just have information come to you as you needed it.”

- Sergey Brin, Google

That quote says a lot about where the future of Google is going.

Illuminated by a brilliant presentation by Dr. Pete, Google is quickly moving towards a world where users rarely, if ever, have to click through on organic search results.

We’re seeing the emergence of local carousels, knowledge graph and aggregated information. We’re seeing the kinds of SERP results that allow users to compare insurance companies, buy newly released albums or make reservations without ever leaving Google.

We’re watching as apps and voice-activated search steal (albeit incremental right now) traffic from search engines, removing the need for a search entirely.

In the future, organic rankings will matter less and less, while organic reach will get harder and harder to pin down. If you want to survive the long-haul, it’s time to abandon some of the familiar, archaic ideas we have and the false beliefs that search will always work the same way.

Tune in. Diversify your metrics. Get closer to other marketing channels, and learn how to add value outside of rankings, keywords and raw traffic numbers.

It’s going to be a bumpy, Exciting ride.