I don't know if it's because I've been living and breathing keyword research tools for the past month, but Bill's tour of Hitwise's features was the most interesting walkthrough I've been on for a product demo. (Okay Bill, I said it--where's my puppy?)
Anyway, Hitwise offers a lot of pretty compelling data. You can generate various reports on a website, an industry, or a search term. In this example, I entered www.amazon.com and then changed the drop-down menu to "search terms" to see which keywords/phrases have brought searches onto their site:
Amazon's search term referrers ending 2/24/07
Clicking on one of these terms (e.g. "books") then shows a detailed analysis of that term for the four weeks ending 2/24/07, including the top ten websites that received the most traffic from the search "books," the most popular search terms containing "books," the search engine share for that query, and the industry traffic share:
Hitwise search term analysis for the term "books"
You can also compare search terms, Google Trends-style, on a chart to see their relative popularity. Here, I compared "prom dresses" (which Bill confessed to having an unnatural obsession with [from a purely search-related perspective, of course...at least I hope so]) to "wedding dresses" over the past month and a half:
Eat taffeta, wedding dresses!
This is all Bill was able to show me, and I didn't get a chance to play around with Hitwise much more than that (stupid Rand and his stupid "tasks"). There's a load of other stuff to check out, such as Search Term Suggestions, Search Term Analysis, and Search Term Gap Analysis, Find Websites by Keyword, Domain Name, or Demographic Composition, and there are plenty of charts to drool over. It would be extremely easy (and dangerous) to get addicted to playing with Hitwise for virtually any sort of competitive analysis comparison.
Now for the cons...as Rand griped, Hitwise doesn't report any hard data--everything is available for a relative comparison. There is no "such and such term had X number of searches, compared to so and so term, which had Y number of searches." This frustrates my Yellow Shoe'd Employer because he knows Hitwise has the data available, so why no sharesies?
A second drawback to Hitwise is that (I'll channel my inner Dave Naylor here) it's fookin' expensive, yeah! Though there's no price listed on their site (they instead ask that you contact them to purchase a package), Rand told me that a year's membership costs around a hefty $20k. That's a lot of money to shell out for no hard data, right?
Though Hitwise costs about half my salary, it's a pretty damn cool program to use for competitive analysis. While it's obviously out of the question for us meager Mozzers (hey, maybe in a couple years we could afford it...), I'm sure that the companies who can afford to shell out the coinage to use it are getting their money's worth.
(By the way, thanks a lot to Bill for taking me through the walkthrough [check out ilovedata.com to read Hitwise's blog], and thanks to Rand for turning me into a stats and charts nerd. If I were in junior high, I'd be giving myself a wedgie right now : / )
MORE FROM RAND: I'd add that we, at SEOmoz, had access to a Hitwise account through a client for about a year (4-5 months ago). I frequently used the data to provide insight into several questions about competitive intelligence, relative keyword volume, search volume and demographics. The data is remarkably valuable and useful if you're a site getting 2-3 million or more visitors each month, and considerably less useful below that (likewise, analyzing competitors smaller than those sizes is also tough).
I wouldn't recommend against buying it, but I would recommend that you identify valuable uses for the data before you make a purchase. I have no doubt that Bill's team would be happy to give you a demo :)
As a frequent user of the Hitwise suite, I'll attempt to clarify a few things that have popped up in this thread:
Why you don't want to use Hitwise
Hitwise is not a keyword research tool. The Hitwise product is a competitive intelligence suite with keyword research included. If you're only after keyword research, Trellian's Keyword Discovery is a much more viable solution.
Sample, not cookies Hitwise uses data taken from ISPs. As such, there is no 'hard' data as they are only working with a (massive) sample of weblogs which they crunch into market share stats. Hitwise will also not tell you which ISPs they take data from. This methodology is different from your Comscore/Nielsen cookie-based services that report unique visitors, page impressions etc.
At-work users underrepresented
As Hitwise only takes info from ISPs, I don't believe they accurately estimate the amount of corporate/at-work usage. For example, the majority of users on our network access our sites at work (especially before work, during lunch, after work). This leads to situations where Hitwise might rank us as #2 whereas Nielsen Netratings shows us a clear #1. Different methodologies sure, but why the big disparity? And guess which measurement system we'll use in our press release?
Why you want to use Hitwise
"The price really shocked me! Ok, I understand the full power of these stats when trying to gain a competitive edge, but is it really worth all that money!"
I've mentioned these previously but here are real world examples of how I've used Hitwise in past weeks:
1. Reporting weekly on our relative market share (RMS) of search engine traffic. I love this metric (and so does management). It provides an accurate representation of your standing against a competitor.
2. Find partners: using Hitwise's Clickstream report we identified a site that sporadically drove 5% of upstream traffic to a competitor. Turns out it was a site with a large email list. We talked to the site and got the same sort of deal, thereby increasing our customers by several thousand.
3. Search term gap analysis: This simply takes all the keyword referrals that generate traffic for a competitor that didn't generate any for you. This functionality is in most advanced keyword tools, but the Hitwise data is very comprehensive. We then sorted these keyword gaps into themes which were subsequently worked into our SEM and then SEO campaigns.
4. Explore spikes in traffic: A competitor in one vertical had a huge traffic surge in terms of visitors and page impressions. We went on Hitwise and found that we could pinpoint the exact day where their traffic spiked (and ours dipped). Diving in further, we found their referrals from search engines and major news aggregators were the major traffic drivers. This led to a renewed focus on news and search optimization, especially in regards to timeliness and accessibility of our own content. To keep a long story short, we knew when, where and how their traffic hit and reported this with a set of charts and tables that even our senior management could understand :)
I currently have access to Hitwise and I'm not sure how much I will miss it when I move agency-side (I'm pretty sure that the place I'm going won't splash thousands on it).
Hitwise definitely has its uses;
For keyword research (finding out what's driving traffic to the competitors)Potential link analysis (finding out who is driving traffic to the competitionMarket share - what slice of the traffic do the competition have
Whilst these are very useful, I would add a few caveats (some of which are similar to those above):
Time - there is a huge amount of data available on Hitwise, but you need the wo/man hours available to make use of it. Otherwise it is a very expensive keyword toolData - As mentioned above the data is from ISPs, and skewed to consumer/home traffic - if you're working in the B2B sector, the data can be hard to find
My biggest gripe (and its with how Hitwise is used, rather than the tool itself) is with regards to market share. We use it to track the market share of our sites, but primarily against the web versions of our traditional offline competition. If you don't know about the little start-up that's sweeping all before it in your sector, it's quite possible that Hitwise won't tell you about it. The industry breakdowns aren't really niche enough, so unless you set up a custom marke share report, and know about all of your competitors, you could be sleepwalking into trouble.
Having said that, all the people at Hitwise UK are very friendly, and Heather Hopkins always has some very interesting things to say about the internet and the various businesses that frequent it...
ciaran - excellent points. And, yes, I too am a huge fan of Heather - she's one of my favorite people.
Isn't she just - lunch with her last year was very amusing & interesting - fascinating insights, delivered with a wry drawl....
Ciaran and Rand, Ah shucks, you guys are great too! Thanks for the really nice comments.
Rand - nice to see you at SES.
Ciaran - hope to see you again before long. Send me an update about the agency you are moving to!
Hi Heather - didn't realise that you were an SEOmozzer too - I'm worried now, as you know that I don't actually look like a cartoon!
The agency is called Eyefall but I'll drop you a line when I start (6 weeks & counting!)
Ciaran, Please do get in touch before you go over to Eyefall. Best of luck! HH
very cool Rebecca, but yeah have to agree "fookin expensive, yeah" That's just absurdly expensive.
I checked out Microsoft adCenter Labs new keyword forecaster (you have probably seen in before) but I thought that was pretty cool.
But I must admit that I usually use PPC campaigns to research keywords, gives you the most accurate data in terms of impressions and you don't need to spend loads of money! Well at least not 20k..ouch...
Hot damn, tried a search on that MSN keyword forecaster for "seo"!
Great review Rebecca! I totally agree that you can get lost in this data, I often do when researching for my client.
Here's something I didn't read in your article, but perhaps should be mentioned.
Hitwise does not sell directly to agencies like SEOMoz or my employer (commerce 360). Trust me, we've tried - and tried and tried.
Our clients each have their own license and then give us an "agency seat". They tie these license seats via cookies and ip's, they also track and limit where you can go and what you can see according to the license.
So - you can't be "evil" with Hitwise, you have a license with your client and they are a toy manufacturer, and that's the industry they subscribe to, you aren't going to see Sports related information on sites in the Sports category. (I believe that'd even be a violation of the TOS and could get your account banned). Also, there's an ethical issue there too - I know myself, I'd never do it, but trying to using someone else's Hitwise account for other data, is just wrong.
Anyway - great summation! :)
Interesting - our (company) licence allows me to look at data for any category, whether or not its related to our own sites....
It all depends on how your license is set up with them. If you have the "full" no holds barred package (uber expensive) then you can look at all channels.
However, to make it affordable to smaller companies, Hitwise sells limited packages, and thus limits your access - the toy company would be the example I gave for this.
they still track agencies very carefully though, to make sure they are adhearing to the license's TOS.
Ahh - we've obviously got the uber-expensive one!
Great, thanks for pointing that out, Li. I didn't know about that...
It is a very cool tool and the price is related to the verticals you want data for. We had it at my last company and only had 2 verticals and it cost us about 30k if I remember correctly. These are jsut some of the tools - their before/after traffic is pretty cool too to see where people are primarily getting to sites and then where they are going afterwards. You also get site demographics as far as who is viewing the site. But like you say - why not give the actual data.
Thanks for the vicarous run through. Does look very cool and I can totally see how you could get sucked into detailed analysis.
Perhaps we could all pool our funds and get one joint account... now, where to get the other $16,000?? Hmmmmm.
Actually, the hefty tag may make sense and I don't think they will take any issue to your comments at all. Much of the value is in the "exclusivity." Those who purchase are not just getting access to a powerful data source, they may be buying competitive advantage over those who aren't choosing or aren't able to buy it.
Even in SEM there is high-end shopping where if you have to ask how much it is, then you can't afford it to begin with.
I was fortunate to have access to Hitwise at my old company. Granted I didn't know the expert uses of it, but found it very helpful on many levels. One of the first recommendations I made when I started my new job was to get on board with Hitwise. They were very open to it until they saw the cost. We all agreed, it's a great product but priced too high for smaller departments/agencies.
Doesn't the puppy count as payment? And doesRebecca have enough time to housebreak a puppy??
I used Hitwise at my last job. (Before I was a poor, happy sole proprietor). IMO, if you have a $1mm + budget it is probably money well spent, but you should make SURE you have the person-power to really use the tool -- it takes time and if you don't really work on it you probably won't get your money's worth. The competitive data is probably the most valuable use of Hitwise. Thanks Rebecca for spending another Sunday blogging!
Hi Rebecca,
(IMO) People should seriously give it a test drive before they subscribe. As you said... no hard data.
Most people who can afford a package like that will usually have a few spikes or dips in traffic because they did something like make the front page of Digg, launch a new product, send out an email blast or Press release, etc...
HitWise should be able to somewhat match these spikes in traffic to your Web Analytics package or not. They won't match exactly but you should test it against available information to see of HitWise shows Mystery Spikes or completely misses the incidents where your traffic tippled for a few days here and there.
Take Rebecca's report above plus Microsoft's report of the same keywords "Prom Dresses" / "Wedding Dresses". I put the Microsoft output chart here so you don't have to go do it yourself. The graphs look completely different for the reported period.
I know Microsoft is a sure bet. I mean those guys produce bloated-blue screenware and everyone believes they are the best.
Wait, I'll visit my clients and show them the graphs and ask if they want to spend $20k with me or with some dumb software.
Anyone got Trellian? I would love to see their chart.
Then there's the Google Trends data, which doesn't look very similar to Hitwise or Microsoft. Makes it hard to know who to trust, I'd say.
I think it is hard to compare though as the Hitwise, at least the chart above, only covers two months. I don't really think it is that far off of Google, when isolating to that period.
The overall trend comparison between MS and Google is similar... the start is a little funky admitedly, but I wonder how much real history they are presenting versus extrapolation on the front end?
Or are these illustrating some demographic differences? My guess is the demographic searching for prom dresses is very heavy Google where the demographic for wedding dresses is much broader and probably represents a more even spread between Google and MSN, as well as Yahoo users.
I'm pretty sure that with Hitwise you can compare keywords over a max. period of two years--I just chose two months for the example.
Yep, 2 years, and i believe you can request even longer periods as a custom report (additional cost of course).
Interesting. So I suppose the question most people are going to want answered is:
That's the data, and it's all good. But is it worth it?
Ignoring the fact that different niches are going to be different, if you were buying a keyword tool, where would you stick your money?
Half your salary? The cost of HitWise pretty much IS my salary. I think I need to get a new job :(
Anyone have any clue how much Comscore or Nielsen charges for similar services? How do their costs compare to their services/accuracy? Anyone else worth comparing as well?
Will you have international keyword results?
Our licence is UK data only - adding the international stuff puts the price up even further!
I tried Hitwise a couple years ago. Some of the data is top notch and well worth it for companies with the budget.
Trellian's Competitive Intelligence service is much more affordable and serves up some great competitive keyword data. For those with only a couple hundred dollars to spend I would recommend it.
I'm pretty sure Trellian is using Hitwise (or another similar service) to get their information anyways. They could be getting it direct from the ISP's but "that aint cheap", which is one reason why Hitwise is so expensive.
We do a brief overview of Trellian's tools in our keyword research guide.
Rebecca...I want to play, too. Could we all get a one day free pass? Are they kidding about the $20,000 a year? What else can it do? That didn't seem like much considering the price. Was he holding back on showing you other stuff?
Hi Rebecca - I have also had a recent demo with Hitwise. I was really impressed with the uniqueness and accuracy of the competitive reports.
The price really shocked me! Ok, I understand the full power of these stats when trying to gain a competitive edge, but is it really worth all that money! I have a number of clients who are interested in the data, but seeing as it will only really be me that is looking at trends and coordinating campaigns, I have considered asked them to split the cost.
The Search term tool got me really excited - As you will know, over here in the UK, it's difficult to get UK only search data. There are a few tools out there, but not as many as you get across the pond. So I was eager to see what Hitwise could offer with their UK specific data.
The lack of hard data was therefore anissue and I was a little disappointed.
I hope to gain full access to this product shortly, I think Hitwise is one of the most powerful analysis solutions available.
...slaps a "kick me" sign on Rebecca’s back and runs off...
Yeah, I figured that any company that can afford to pay that much for Hitwise should be getting some value out of it--hence, they must be getting their money's worth (or at least telling themselves they are).
Hi Househunter,
I'm in Leeds so I feel your pain. WordTracker just added a uk based db and you can select the uk-only results, but they haven't built their db up to that the US based one is yet. It will be nice when they do.
GaryTheScubaGuy
I was talking to WordTracker about this the other day - apparently they were planning to launch the UK db at SES London - when it was due to be held in the summer. When it was brought back to Feb, they were suddenly left a bit in the lurch....
Hi Scuba guy,
I think there is something going o with UK database in wordtracker. Results just don't seem right! Do you get the same?
I'm in Sheffield btw, send us a contact on skype as you mentioned you wanted more qualified staff in the UK ;-)
Like I said, from my conversations with them it appears that their UK database isn't up to scratch yet as they had to launch it earlier than planned due to the changed dates of SES London. Give them a call if you've got any questions - they're very helpful.
Where is Hitwise getting 2007 Google Trends data? Google sure isn't showing it. I even manipulated the URL and I recieved an error. Is it 2006 data, an API or does Google have a special relationship with Bill? ;)
I said the trends charts were "Google Trends-style," meaning similar. It's not directly their data, though.
Ah, got it. Never mind. My bad. Moving on.
Yes, I'd like to know where is Hitwise getting their data from.
Is SEOmoz doing paid reviews now?
Kurt - We're certainly not, and if we did, we would, of course, let you know about it in the post.
Not a paid review--I just thought it was one of the more interesting keyword research/competitive analysis tools I've been researching for the past month.
Not sure if Kurt was being funny or not, but I think its interesting how everytime a site does a review now (with paid reviews being such a hot subject), people feel compelled to ask if it is paid or not.
If this "was" a paid review it would be a great example of how to do it right. It is well balanced and open to public comments. Hell, I even linked to one of their competitors :). We could only hope that all reviews, paid or not, were as well written.
With that being said I dont know any professional bloggers that do not disclose when a review is paid. Rand "Transparency" Fishkin, or his employees, would be the last people on the planet to get paid for a review without disclosing it.
I thought the same thing Kurt did, but I didn't want to upset anyone, so I did the polite thing and didn't ask. I think Kurt's question raised an honest concern, and I'm glad he asked it.
I believe the answers Rebecca and Rand gave, but I'm also glad the issue was raised. It cleared away my lingering thought.
Christina
Hi Tinkerbell,
It's good that the question got asked, especially if there are other people wondering.
I just find it really interesting how blog readers are reacting to reviews now. It's directly related to the launch of all the Paid Review Services.
The interesting thing is that reviews are a core part of the blogosphere. Anytime a blog mentions another blog or company, it is giving a review. It was only recently that people started publicly questioning whether a review was legitimate or not.
It wont be long until every blog has a disclosure policy linked to the bottom of every post, to curb some of the questions.
I think this is a beautiful thing in many ways. The attention that paid reviews are receiving is forcing bloggers to pay attention to how their readers perceive them.
Many novice bloggers do not realize how important keeping their audience's trust is.
Most of the Pro Bloggers know that if you are going to maintain trust, you need to be completely tranparent.
Anyways, sorry for going off topic. SponsoredReviews has been my life for the last 6 months :), so I track the topic heavily.
Rand, any thoughts of putting a disclosure policy up? It might put some readers minds to rest, and I think it does a lot to build trust with new users/readers. Readers that might not already know how credible SEOmoz is.
Excellent idea, Jarrod - I agree. We'll try to get something up in the near future.
I can understand how one might think it was a paid review, but as Rand stated above a disclaimer would have been in the post.
I think this is going to be a major problem in the blogosphere in the next few months/year.
Rebecca - I am not so sure it is good business to come out with a "keyword research guide" and point out that someone elses is way too expensive at the same time.
This post doesn't appear to be "paid" but it surely can be taken the wrong way, not sure if I worked for hitwise I would be happy.
Bill seemed fine with the write up. Besides, we're not talking about the Competitive Intelligence tool in our guide--we're talking about the much less expensive Keyword Intelligence tool.
...and even if I were to say that one tool is way more expensive than others, it's only to benefit our users. The guide's not for the companies who provide the tool, it's to educate people who may use the tools.