August 2008


27 Aug 2008 12:34 am

vortex Save the best for last! Insights for Search is one of the most awesome research tools you'll use. I promise.

We have learned how to use Trends for Websites to measure core site metrics and identify strengths and weakness of our competitors.

We built on that by learning how the Google Ad Planner can help us identify demographic and psychographic attributes of the Visitors to our website, as well as our competitors. Not willing to stop there we also identified specifically how to identify Publishers where our target audiences exist, to aid in our intelligent acquisition efforts.

In the final post in this series we cover Google Insights from Search and go back to good old keywords and mining search behavior by our customers to find actionable insights.

Given the dominance of Search as a acquisition channel, you are going to have fun here.

Do the obvious thing first: Understand "head" keyword trends.

Type in a keyword and check out the trends over time, a quick peek into popularity of the terms you are interested in. . . .

Step One: Type keywords. [I am going to compare volumes for my friend Bryan Eisenberg - Always Be Testing -, Garr Reynolds - Presentation Zen - and myself to see if I can learn anything.]

google insights for search queries

Step Two: View Data.

google insights for search trends 1

Some quick thoughts.

Bryan's search traffic has stayed pretty consistent over time, steady over the years even as the "noise" in his unique space, persuasion architecture, has gone up quite a bit. There is a slight lower slope in the last 12 months but overall Bryan has great strength.

Garr shows up on the scene in late 2006 and you can see him stay steady and then in the latter part of last year he announces his book and you can see his plans for world domination take off like a rocket, over Q4 and he is not looking back (except in the last month).

[What do the numbers mean: The numbers on the graph reflect how many searches have been done for a particular term, relative to the total number of searches done on Google over time. More here.]

Remember always look at your data in context, hence the slope of a individual line is important but it is in context to the other lines that it gets to be more fun.

What can I do with this?

Run the graph above, atleast for all your head search terms (top 20 terms that bring most of the traffic) and see if your indexed performance (against competiton) is good or bad.

Notice those three numbers on the graph above (74, 12, 54)? You can even report monthly performance in your Management Dashboards.

You can also use these trends to validate your business strategy, unaided (or aided) brand recall, customer intent. (Or feed your paranoia: Do you matter!)

People how know me know that I love my ThinkPad (greatest laptops in the world!). You also know it used to be owned by IBM and is now Lenovo.

So I wonder how the ThinkPad brand is holding up over time vs the Lenovo brand. . . .

insights for search lenovo vs thinkpad

Good news on the Lenovo front. From no where the brand has continued to grow strong, as reflected by customer searches, and on a great slope.

Potentially not so great news on the ThinkPad front (me so sad). The brand has continued to erode steadily over time.

This should potentially be worrying because during the same time the business brands of HP and Dell laptops tend to hold steady in terms of search volume (though Dell's seem to on a slightly over time).

Remember all this data can do is raise questions. It could be that Lenovo is being emphasized over Thinkpad, it could be that the competiton has just stayed steady on message with their brands, it could be a reflection of Lenovo's search strategy, it could be that Lenovo believes that its core direct business strategy is so strong that the web one is not quite as important.

So what's wrong? That's for David to answer (in private I am sure :).

You can also use this kind of data (from Google or another source) to measure what kind of brand equity you have built over time for your core business brands.

Here's a example: I was embarrassed that I did not know Lenovo did desktops, I recently ran into the ThinkCentre PC's ("The ultimate business desktop") on their website.

So I wanted to know if I was the only person in the world who did not know this. Here's the data. . . .

insights for search lenovo vs thinkpad vs thinkcentre

Either by design or by fate, it seems that very few people have heard of ThinkCentre PC's – as measured by people actually typing in either ThinkCentre or Think Centre into www.google.com. That inspite of the fact, you'll note the orange alphabet markers above, Lenovo has been pumping out press releases and news about the product.

That's what you can do: interesting analysis about your core head terms. Analysis that helps you index your performance and raise troubling questions.

You can also get a bonus based on this data, for example Seth (who runs search strategy for the green line) probably included this in his annual performance review. . . .

insights for search turbotax hrblock

: )

Google Insights for Search offers a large free source for all this rich search data, but even if you use paid solutions from Compete or HitWise you should be thinking in terms of similar analyses.

Kick it up a notch: Exploit geographic customer demand trends.

Comparing trends over the last four years is just one part of what you can do with Insights for Search. The other large component is your ability to understand them from a worldwide perspective and along geographic segments.

Continuing with the Lenovo theme, let's try to understand opportunities for the X300 (which I really want!).

I start by trying to understand customer search behavior across different countries for the term lightweight laptop (the blue line is for "ultraportable laptop", not quite as popular). . . .

insights for search lightweight laptop

Amazingly the US comes in a bit behind the UK in terms of the search index, an obvious market to exploit. But what about rest of the world?

Here you go. . . .

insights for search lightweight laptop worldwide

Sweet. Three countries stand head and shoulders above the rest, as is obvious. It is surprising to me that Japan is hot much more obvious given their obvious propensity for small stuff.

Also notice Philippines and Malaysia show up nicely.

If I were running Lenovo's search campaigns here is good data for me to profit from: The demand is there, are we showing up in Organic and Paid Search results for these terms?

But it gets better.

You can actually drill down to individual country level and then state level to further understand search behavior.

For that I tried something different.

While the Macbook Air is not a direct competitor to X300 (you see there are Mac people and then there are PC people :) it is possible for us consider understanding its popularity and try to be aggressive and provide the X300 as an option to those folks.

[It is to be noted, perhaps unsurprisingly, that way more people look for the Macbook Air than the X300.]

Here's the geographic search volume index map for the X300. . . .

geo x300

Significant search volume (indexed) on the coasts, Texas and NYC are strong (and I wonder why North Carolina is strong :)).

So how does that compare to the demand for light weight computing as represented by the Macbook Air?

geo macbook

You see strengths in the same states as the X300, but there is pretty nice demand across the country, all states (except Wyoming, what's up with people in Wyoming?).

So clearly there is demand, but people are not looking for us. Clearly great opportunities for a aggressive Marketer at Lenovo to exploit this difference and build demand for the X300.

She/He can obviously use Organic or Paid Search on Google. But in those empty states up top for the X300 the Marketer could just as well consider doing TV or Radio or Print or Newspaper or even buy geo targeted banner ads on www.yahoo.com to get the brand out there and get in the face of those people who are looking for light weight computers.

Bottomline: Insights for search exposes customer demand across geography, you can use that data to optimize your online and offline campaigns to benefit from that demand.

It bears saying that this is not some phone survey data by bugging people at dinner time, this is not panel data extrapolated to infinity, these are not online surveys. These are actual queries typed by actual people who are actually looking for you (or your competitors). You can see how it is vastly superior.

Go for gold: Ecosystem: Identify "related terms", "rising terms".

As all good SEO girls and boys know it is not all just about the terms you are interested in, it is all also about the ecosystem related terms if you really want to win the search game big (on any search engine online).

I have stressed quite a bit on this blog about moving beyond the top 10 or top 20 (tip of the iceberg) to what's happening in the massive long list of keywords (below the tip of the iceberg, hidden under the water). Statistical techniques like "what's changed" are fantastic at this. I am quite pleased that Insights for Search offers a "what's changed" view as well!

Since we were just working on understanding lightweight laptops, here is the first nice bit of insight from the tool. . . .

lightweight laptops related terms

Two action items.

1) You can ensure, if you are doing SEM, that you are taking into consideration all these related terms to ensure your bidding (or even SEO) portfolio contains the right terms.

2) Every time I find competitors here, and not myself, I feel a bit sad. Clearly in this category Users are typing competitor terms and that is sad (Ok, ok I am religious. Ok, I am happy for them. There. Happy now?)

You can see how this can give you insights across other categories, and customer preferences. Remember our use of Pampers as and example last week in the Ad Planner post? So how is the picture for diapers?

diapers related terms

Someone at P&G should start worrying that cloth diapers is so huge in this category. You can easily see how the numbers above reflect the individual success of Pampers and Huggies (the two big brands here in the US).

Related terms are good to understand. More awesomely awesome are "rising searches", terms that are not quite yet in the top but are rapidly becoming important in this category.

Here's the story for our lightweight laptop category. . . .

lightweight laptops rising terms

Whatever it is that Toshiba and HP are doing in terms of their online and offline campaigns (branding or otherwise) is clearly working because terms with their individual brand names are rapidly rising in customer conscious (and hence them typing into search engines). I

Beyond that at the minimum what this list offers are keyword and key phrases that you might not have considered but you should.

Rising searches to me is also a fantastic way to understand trends that will become mainstream in the future, but currently are just rising. So not only do people want lightweight laptops, they also want them to be small. As an example.

So what about our diaper category?

diapers rising terms

Someone at bumGenius deserves a raise, it is astonishing how much they dominate the tidal wave of rising attention at the moment. Clearly the environmental movement is catching on.

None of the main brands are present there.

Perhaps P&G should buy bumGenius to preserve future revenues. :)

I have to also admit that I had never heard of any of the diaper brands above, a reflection of how some of these trends are so hidden under the radar and then end up surprising Marketers when they hit mainstream. By then it is too late.

Bottomline: Use the data above to identify bleeding edge trends for your industry, key words, ecosystem before it is too late.

Some last goodies: Time Range Trends, Locations & Categories.

Couple more fun things you can do with Insights for Search. You can more view trends for different time periods.

I am going to check the performance of my brand over two years and if I am getting better or worse. . . .

insights for search time range comparisons

Click Search and. . . .

insights for search time range comparisons avinash kaushik

No presence in the Jan 2007 time frame (why! :)), and doing much better since then. Doing much better year over year. Happy birthday to me. :)

It is interesting to see the dips over summer. I wonder what causes that (any guesses?).

Do this graph for all your core brand terms, for your category terms, for your performance for your industry "rising terms". You'll have insights up the wazoo.

Another example. A real business. From my visit to P&G, year over year performance for one of their biggest worldwide brands (one of their biggest advertising spenders offline) Gillette. . . .

insights for search time range comparisons gillette

Quite sad.

So much money spent hiring Tiger Woods, Roger Federer and Thierry Henry, so many very nice tv commercials, and so many nice print ads. But online they are doing much worse year over year.

I bet for the cost of airing just five of those TV ads Gillette could easily reverse the trend above, if it really set its heart on winning online.

Switching gears.

You can also do category searches in insights for search.

So what are the terms that are top in the Auto Insurance category in the US?

insights for search location category

And boom!

insights for search location category auto insurance

Delightful, right?

And don't forget to look at the Rising Terms, that's where the real meat is. Motorcycle insurance is rising, travelers and nationwide are both rising brands (good for them!).

Finally you can also do some fun stuff.

I did a search on the "hottest" (rising searches) celebrities (category: entertainment, sub category: celebrities) in New York in the last 30 days to see how much I am in sync with popular culture. . . .

celebrities in new york

Who in the world is Kim Kardashian? I feel lame, I am so out of touch.

Oh here are the ones getting hot in the last 30 days. . . .

rising celebrities in new york

Go Tom Brady and Aishwarya Rai!! And some things are evergreen, go Ms. Anderson! : )

Do I need to explain how you can use this data for business purposes? Think hottest hotels or business schools or computer software or web hosting providers or charitable causes or . . . . well you get the idea.

Agree, pretty damn cool and actionable?

I hope that your have learned how to leverage Insights for Search for your business, but more than that I hope how you can use the techniques in any tool that provides search data (like Compete or HitWise).

Before we close, some helpful links:

Ok your turn now.

What do you think? Was it fun? Did you learn anything new? Did the business cases I used make sense? Have you used the tool, what did you not like about it? Anything obvious that I missed about how to use it? Do you use other tools to get this type of data?

Please share your feedback, ideas, and critique.

Thanks.

PS:
Couple other related posts you might find interesting:

18 Aug 2008 01:34 am

pollen 1Google's Ad Planner is less a competitive intelligence tool and more a tool that gives fantastic insights into understanding behavior of visitors to your website in context of the broader ecosystem.

It also helps answer critical questions that have thus far stymied Online Marketers of all shapes and sizes (especially those in charge of acquisition).

For example, ever wondered not just how many Unique Visitors your competitor got but also if there is any overlap between their visitors and those that visit your website?

Or what are the demographic and psychographic attributes of those visitors?

Or you need to boost revenue by 900% in one week and what websites should you target in terms of purchasing ads so you can find the right audience and get those relevant conversions?

Or compute the amount of potential exposure your ads can get across a complex media plan (which you have created and saved for future reference in a free tool)?

If the answer to any of the above questions is yes then my dear dear friend the missing ingredient from your life is the Google Ad Planner! :)

This is the second post in our recent Competitive Intelligence series. The first one was on Google Trends for Websites , check that out for awesome true CI insights.

As in that case my hope here is to share with you the unique data and functionality that the Ad Planner provides, but more than that I hope to teach you how to think uniquely.

google ad planner tool

The Basics: Research Your Website / Your Competitor's.

So this is the nice stuff, click on the Begin Research button in the Ad Planner and you are all set to go when it comes to understanding the basics of what the tool can offer.

[sidenote]
Google's Ad Planner is currently in a Beta program and you have to apply to get access to it. The wonderful folks on the team tell me that access is provided to anyone who applies, and the process takes just a few days. This week they'll try to be extra diligent and get everyone in as soon as they can – and we will send warm hugs to them for helping readers of this blog.
[/sidenote]

In the left nav type in the name of the website and hit Add and boom (!) you have your first taste of data representing metrics for the last 30 days. . . .

google ad planner omniture

Here's how to read that data:

  • www.omniture.com was visited by 150k Unique Visitors who were from the United States
  • This audience represented a 0.1% reach of the US audience (not bad for web analytics!)
  • This audience also represents consumption of approximately 210 million page views from all their browsing behavior (obviously not just on omniture.com, though I am sure that would make Omniture very happy! :)).

In the right frame of the page (which is a lot more useful later in your analysis) the data is represented thusly. . . .

google ad planner omniture 2

Now you also have a feel for Omniture's page views there.

Next up the wonderful feature where you can get a much better understanding of the demographic and psychographic attributes of the Visitors to Omniture's website. Simply click on the icon above next to the website's name, and you get this. . . .

omniture website audience characteristics

So these kinds of people like to pay lots for web analytics. :^) Just kidding!

One important thing to note is that it would cost you lots of money to get this kind of data with any degree of confidence, it is now free. So use it.

The wonderful thing about this is you can start to form the basics of a persona of the kind of people who visit Omniture's website. I was for example really surprised about how many were there from the lower household income ranges.

Perhaps they are there to learn from Omniture (validating Mr. Funk's strategy of having quality educational content on the website).

Perhaps it also indicates that people who actually sign large sized chq's could not be bothered with learning anything themselves. Zing! : )

But it is in the contrast that this data shines.

For example visitors to coremetrics.com are a lot more likely to be highly educated (by almost 18%) and with far far higher incomes (by atleast 20%)! That can give you a pause, if you owned omniture.com, and see if your marketing programs are actually driving the right kind of traffic to your website. The hypothesis being that people with higher income profiles and higher education will, supposedly, cut bigger chq's!

Or why is it that when almost every good Web Analyst I know is a female that there is a grand canyon sized gap between the male-female ratio for www.webtrends.com? Wrong targeting?

You catch my drift about all the things you can do, remember: contrast gives context which gives insights.

Finally you can compare and contrast site traffic data quite easily, just punch 'em into the left navigation and boom!

site traffic comparison coremetrics webtrends kaushiknet

Guess which one is CoreMetrics, Kaushik.net and WebTrends?

Yeah baby, www.kaushik.net/avinash, is #2! [Sure a little braggy, but you will indulge me just this once right?]

WebTrends is at 84k worldwide in the last 30 days (I just need another 20k! :)). To get the worldwide data you'll simply green "graph icon" next to the site's name in the results.

Again food for thought if you are running those websites and want to get context to your performance against a random blogger.

[sidenote]
Quick update post post posting :), here are links about Ad Planner data:

[/sidenote]

The Advanced: Research Ecosystem, Customer Behavior, "Related Sites".

One of the core drivers behind creation of the Ad Planner was to:

Empower Advertisers and Marketers to identify related websites and audiences that they are interested in.

Here is an example, I wanted to compare my own performance, for www.kaushik.net, with the various web analytics vendors, with a eye towards understanding audience differences. Here's what it looks like [for the US traffic]. . . .

google ad planner site traffic comparisons

Notice an interesting thing that the Google Ad Planner Tool does. When I typed in the first five websites listed above the Ad Planner it will also bring forth "related websites" based on the audience persona (or technically: demographic and psychographic attributes). The last three above.

So even as you go about your merry business the delightful tool will highlight to you others you might not be thinking of, WAA for example. They were not top of mind, but now I see them there and I can dig in and compare traffic stats and the types of personas etc.

The website's you notice are sorted by Comp Index. Here's the definition: The composition index shows how concentrated your audience is on a specific site relative to internet users within the country you have specified.

The higher the number the more overlap in the kind of audience you have chosen.

What the analysis above shows is that audiences who visit kaushik.net are significantly likely to visit the web analytics vendors websites (indextools especially, Dennis you owe me something!). Those visiting the WAA are likely as well, but a lot less.

Let's apply this to a real world example, after a while web analytics is a boring. :)

I recently had the pleasure of speaking to a august audience at the P&G headquarters in lovely Cincinnati. Wanting to brush up on them a bit (and what better way to do that then use CI tools – free tip for you all), I used the Ad Planner to understand www.pampers.com.

If you type that in you get this nice bit of data. . . .

google ad planner pampers

You get a very fast snapshot of the core metrics of the Pampers website. But more than that I also get a really wonderful understanding of the websites that its audience visit and just looking from the names I can get a feel for the persona of that audience (free stuff 4 baby, freebiecountry :).

The next thing I did was to click on that "green graph" icon next to the site's name and get the demographic attributes of the audience.

Of course waaaay lop sided towards females (what is up with that, I made all the diaper decisions for our family!). But I was not sure of the context so I compared that data with the one for huggies (the biggest competitor).

comparison pampers and huggies sm

It turns out that Pampers gets a lot more traffic than the main huggies website (which seriously could use the touch of a decently qualified UI and UX person), but the demographics show that the persona of the visitors to the huggies website is that of slightly younger women who might have slightly less education and have a slightly less household income.

Interesting.

If I am recalling correctly that is also an audience where the birth rate is higher and growing per the latest US census. So as the highly paid Brand Manager of Pampers I would like to be in front of that audience, and possibly sell lots more diapers.

So what to do? Throw up more tv ads and indulge in other "faith based initiatives"?

Nah!

Go back into the ad planner and create a media plan that you can hold vastly more accountable.

I go back and choose the attributes of the audience that I want to target with my advertising. . . .

google ad planner media plan

And as soon as I hit that last check box I get a list of websites that are visited by the customer profile I have created (approximated 13 million Unique Visitors in the US!).

Perfect for me both in terms of figuring out the unique persona I am interested in. . . .

google ad planner media plan target websites

Nice ain't it?

I understand my website better as a part of the ecosystem. I understand a specific competitor and their strengths. I can then go in and find the audience that is their strength. All for free. :)

The Awesome: Identify Advertising Available, Create Your Media Plan.

The Ad Planner will not only help you identify the optimal targets for your advertising opportunities (along with key data such as unique visitors and all that nice candy), it also helps you get key data, where available, about options for actual ads you can run.

For the websites above (perfect for the Brand Manager for Pampers). . . .

google ad planner media plan ad options

The last two columns show if advertising is available on those sites through Google and if so then approximately how many impressions and what kinds of ad formats. Text ads, image ads, video ads and gadget ads.

For many sites Google does not have ads to sell (like on cafemom.com or forever21.com above). Nonetheless this data is still very helpful to you in then going and buying that ad inventory elsewhere, while silently thanking Uncle Google for helping you get access to this delightful data! :)

And it goes without saying that you can save your media plan, you can create as many of them as you would like to (say you are a advertising agency with many clients) and yes you can even export all this data into Excel.

This post started by mentioning that the Google Ad Planner product is less a competitive intelligence tool and more a tool that provides you free access to some absolutely delightful demographic and psychographic data that is useful and actionable. Hopefully by walking through the cases above you'll see exactly how it does that.

Regardless of your use of the Ad Planner I hope that you'll be a lot more aware of the data that is available to you and exactly how you could use it.

Finally, if you want to learn more about the "unreliable world of online marketing and analytics" then might I recommend my good friend Ian Thomas's blog Lies, Damned Lies… He has recently switched from a Analytics role at Microsoft to a Advertising role, I think you'll appreciate his insights and, perhaps even more, his distinct sense of humor.

Ok your turn now.

What do you think of all this? Surprised to find it on a "web analytics blog"? No? Yes? Why? Have you used the Ad Planner? What do you think of it? Tried some other tools? What do you think of them? Notice my distinct lack of encouragement for you to use demographic data in a silo and just on your site?

Please share your feedback, ideas, critique and love.

Thanks.

PS:
Couple other related posts you might find interesting:

Ad Planner Help:

Next Page »