Guest: Marty Weintraub with @aimclear. Marty wrote a post on #WTF Factor: #SEO Analytics Survival In The Age Of Vanishing Keywords we’ll be discussing in detail. Marty is author of “Killer Facebook Ads” (http://t.co/4fnNsAag) & CEO of @aimClear, an online search & social marketing agency.
Check out @aimclear ‘s blog: http://t.co/CYGpmKmd – great stuff in there. Marty speaks regularly at conferences like MediaPost Search Insider Summit & OMS, Pubcon, SMX, and many others. Marty writes for industry publications such as SearchEngineWatch, SearchEngineLand, SearchEngineRoundTable, and the aimClear Blog.
Is (not provided) about privacy, Adwords, or something else?
There’s no actual privacy argument for masking this organic KW data, though. You can buy it just fine using AdWords, get your check book out everybody. If you want to test your organic funnels, and truly understand the KWs, then you now have to buy AdWords for empirical testing. AdWords PPC passes KW data just fine. So what gives? Does Google hate SEOs?
To our minds, some missing KWs there to be had – with a little ingenuity – BUT organic conversion @ the KW lvl is tricky. Killer hacks are there, ex: mashing conversions @ pg lvl w/ good semantic insight from GA & Webmaster tools.
@ashbuckles: Is there an industry you’ve found that doesn’t recover keywords well? Or just occasionally run across them?
@aimclear: Have not found a correlation at the vertical level, to what sites can recover data.
@JadedTLC: I’d argue technology sites <- sites with high google plus adopters.
@beyondcontent: I do mostly big brand fashion SEO and most clients now have (not provided) in Top 5 organic KWs which is concerning.
No reason to panic. The sky has not fallen with KWs not provided. Some sites can recover 10% – 30% of missing KW data, if you know where to look. SOME sites. Using webmaster tools, Excel, Google Analytics, and some serious elbow grease. AND, there are so many other metrics, that are meaningful, other than granular KWs.
First, we have not always had KWs, Remeber AOL?
I have a question #SEOChat, Do you think this matters? Do you hate Google for it? What metrics do you use now, THEN I’ll share my hack.
@scotttdodge: I certainly think it matters, and it’s a bit frustrating. I tend to make an educated guess based on landing page.
@Josepf: It does matter as an annoyance, but not really because can smooth it out w/good baseline statistics.
@OptimizePrime: The best way I’ve found is to replace the np keywords with the actual pages they land on. product page vs home is helpful.
@iPullRank: [Not provided] doesn’t matter to me. I track my audience and performance of landing pages I know my target kws per page.
@aimclear: Agreed, but did you know you can recover KWs at the page level, using Webaster tools?
The technique is to work at the page level, in Webmaster tools, you can look at search KWs at either the site or page level. There is little to be gained at the site level.. So here is what you do.
- First, go to Google analytics. Pull KW data at the individual page level, put the list on a spreadsheet.
- THEN, in webmaster tools (WT) pull the keywords at the page level as well.
- Match the common words in WT with GA. You will see that WT, for the few words that it shows data for, OVERREPORTS common words sometimes http://t.co/rQEVWGHs . If you look at every page in your site, in both GA and WT, find the words that WT overreports, then you can often recover KWs.
Trouble is, this is all about the short tail. Google goes to great length to mask the long tail. They don’t give KW data for the long tail. HOWEVER there is another way to infer the long tail KWs that may be driving traffic. Yep, data not very tight, 34.6 inch yardstick better than NO yardstick at all, that’s for sure .
WT Does give some data, about average position, correlate average rank in WT to Click likelihood. The chart I uploaded shows the position of misc KWs that WT ONLY gives position data for http://t.co/v7r5kDNt . You can see that (obviously) higher ranking KWs in WT tend to click through. SOOOO….
If you chart the KWs at page level in WT that have high ranking, you can plot the certainty that they may be missing long tail.
Another thing – Google Analytics sees [Not PRovided] in Custom Segments. GA shows KW not provided in custom segments, check conversion on ALL the missing KWs 🙂 LOL http://t.co/tEJH75dJ .
The good news WHO CARES, there are so many other ways to track SEO Success. I do this only as mental fodder. Track sitewide CTR, and filtered by KW group concepts in Webaster Tools, capture the data each day, create trends over time.
Bing works well for KWs still. Bing’s volume isn’t as great, but Microsoft also isn’t masking organic KW data. That’s a plus! Also, at the end of the day, TOTAL organic performance often, and has always, mattered as much as granular, right?
Clients also understand bucketed organic conversions from new visitors. KW diversity, organic impression count & rising or falling FTR are valuable. Why? New Vs. returning visitors matter. Most clients understand when SEO efforts yield new visotors from organic KWs, even if we can’t see the KWs!
GASP, Unpersonalized Ranking Reports! SEOmoz & Raven ranking report tools are cool Check ’em out. Unpersonalized ranking reports provide general glimpses of how pgs perform @ the KW level.
@JadedTLC: I agree. You look for trends and you still have 70-80% exposed keywords; you can assume similar experiences.
Another important point, know the difference between not provided and NOT SET. Not set seems to be all the shit that Google can’t keep track of.
@RicDragon: Curious – successful in NOT having to give serp reports to clients.
@aimclear: The same dumbass clients who wanted ranking reports, still want them, others follow our advice, let go
@dohertyjf: IMO the hardest thing about [not provided] is year over year metrics, for granular stuff. Big metrics are untouched. But for clients that we were reporting on unbranded kw’s as a metric, now that’s gone YOY. Oh well. Report on other metrics. Other KPIs, after all, are more important. Like profit from organic channels. Once expectations were reset, was able to shift reporting to overall organic traffic. Unbranded is a ballpark.
@ashbuckles: Ranking reports can be valuable at certain stages of the campaign but overall they’re inaccurate and distracting at best.
@aimclear: yep, we’re pretty much hosed in Google. Our advice, use the data you have, buck up, forget it and move along. That’s why KNP in GA segments are so important.
@RicDragon: Prob be cool if you could say out of X amount of searches on topic of “blue widgets” you ranked xx
@dohertyjf: As @RavenJon said “Rankings? Have I increased your traffic? Ok then!”
@JadedTLC: You can also still get profit numbers from organic (or traffic for content sites) and that alone grows, you show success.
@dculmone: Get a lot of requests for conversion metrics, but this differs so much from company to company and project to project.
@dohertyjf: IMO, np is harder for those of us trying to create useful content on sites that people are looking for already. So it hurts the quality of content. I have a client w/ 35% NP. That hurts. No longtail! Also, GWT can be a relative metric to use for SEO success. Cant break into brand/non, but overall it may work.
@beyondcontent: Which is what @seobook blogged about most recently. http://t.co/CB8a3u93 .
Webmaster Filters can also allow you to create “concept groups” of KWs to track.
@scotttdodge: I compiled @aimclear’s knowledge for y’all. Google doc here: http://t.co/nWfdSTID .
@ashbuckles: Hashtracking shows 510 tweets, 2MM impressions, 200+ followers within the past 24 hours: http://t.co/cus768fg .