CMS Watch Web Analytics Feed

Migrating from HBX to Omniture

philkemelor@pkwc.com(Phil Kemelor) — Tue, 24 Jun 2008 00:50:00 -0400

Web Analytics Report readers know that one of the biggest issues about Omniture SiteCatalyst is the complexity of the implementation. Former HBX customers (now part of the Omniture stable) are finding that SiteCatalyst is a significantly different tool because of the amount of customizations required.

Migration becomes a double-edges sword: you can get more functionality, but you have to work harder at it, and you may lose out on some of the things you liked about HBX, such as tagless campaigns. Although I hear reports of HBX customers leaving Omniture forCoremetrics' and WebTrends, there still seems to be a lot of interest in simply staying the course and enduring the migration.

As I've argued before, if you have a contract coming up for renewal, use this as an opportunity to check your requirements and test the waters.

However, if you want to go for the migration and you'd like to get a vendor-neutral take on how to deal with Omniture's ins and outs, I recommend taking a look at the Implementation Toolkit for SiteCatalyst published by Semphonic, the firm I work with. It's comprised of 6 parts: Implementation, Best Practices, Migrating from HBX to SiteCatalyst, SiteCatalyst Coding Standards for Flash, AJAX and DHTML, a Project Plan template and Functional Spec template. It's written for web analytics managers who are actively involved in HBX to SiteCatalyst migrations, SiteCatalyst administrators, and CMS managers and application developers who need to make sure that data is collected properly for Omniture SiteCatalyst.

As always, whether you plan to migrate or are considering another option, make sure you have figured out not only your metrics requirements, but whether stakeholders are actually using the data you're providing. I see too many situations where reports are simply thrown over the fence to stakeholders and they have to figure it all out on their own -- a waste of time and resources, whether you're using a free solution, like Google Analytics or Omniture.

The challenge of mobile analytics - Part 2

philkemelor@pkwc.com(Phil Kemelor) — Thu, 12 Jun 2008 06:10:00 -0400

For those who've read The Challenge of Mobile Analytics - Part 1 and hoping that the picture is brighter in Part 2, well...it all depends.

If you are hoping for a silver bullet solution that will give you razor-sharp accuracy, and provide more robust metrics than the "traditional" vendors, you may be disappointed.

To recap: some new analytics vendors market themselves focused on mobile analytics, both from a branding perspective and because they have clearly committed to figuring out how best to capture and report on mobile web data. Meanwhile, traditional online analytics vendors haven't been as aggressive in this area. However, this doesn't mean that the mobile analytics vendors have necessarily come up with great solutions either. Right now, it seems the mobile-oriented players exceed the online analytics vendors in their integration of WURFL data as part of the service or software, usually enhanced with DeviceAtlas (which is considered to have a more current database than WURFL for phone and manufacturer information).

Let's look at a sampling of mobile-oriented vendors.

Amethon is a real-time reporting solution that has its roots in Australia and also provides multimedia (MMS) and short messaging services (SMS) tracking. The company's primary selling point is that it uses a packet sniffer-based (server side) data collection model. While this makes sense as a way to get around the many client-side challenges here (such as lack of mobile acceptance of JavaScript), if you don't want to host your own analytics solution, or you face resistance from your network admins in bringing new hardware into their environment, this could be a gating factor.

Bango Analytics is a UK-based SaaS offering that is an adjunct service to Bango's mobile content pay solution. Focused on campaign tracking, Bango Analytics uses an image tag data collection methodology based on directing campaign respondents to a Bango network landing page. This could be a constraint if you don't want to send people to a Bango URL.

Mobilytics is a SaaS supplier that also uses an image tag to collect data. However, the company says that they will incorporate a packet sniffing option as well in the next release. Mobilytics also claims that it will apply a scoring model and slider function to present a visualization of unique visitor accuracy based on the source of data collection. Following the example set by Dennis Mortenson to assess IndexTool's ability to meet WAA Standards, CEO Greg Harris recently provided some insight to how Mobilytics compares. A good effort, and one that should help add transparency to this new analytics field.

I'd expect the mobile analytics vendors to aggressively differentiate themselves while the online analytics vendors get their act together within the next two- to four years; but you'll still need to carefully assess whether their methodologies, architectures, and resulting reports make sense for you.

CMS Watch Competition Winner

aps@cmswatch.com(Alan Pelz-Sharpe) — Mon, 2 Jun 2008 09:45:00 -0400

You may remember a while back we launched our little competition to come up with a new name for our vendor positioning chart. We had some great (and varied) responses from all over the world. And it took quite an internal debate to decide on the eventual winner, but decide we did. And the winner is...Laurence Hart who offered us the name "Cross Check." Laurence, a bottle of champagne is yours.

Over the next month we will continue working with our designer to revamp the chart, and of course to rename it -- so look out for the Cross Check in all our report updates this year.

The challenge of mobile analytics - Part 1

philkemelor@pkwc.com(Phil Kemelor) — Thu, 29 May 2008 00:06:00 -0400

Last year everyone was talking about Web 2.0; this year it's all about the mobile web. Let's take a look at what this means for mobile analytics...which by the way, I'll be speaking about on a panel at the Internet Marketing Conference in New York on June 4.

The bottom line...mobile analytics is relatively new; beyond infancy, but certainly not for the faint of heart.

Biggest challenges are

Identifying unique visitors, because mobile browsers don't often accept cookies (you will likely need to rely on browser, OS and IP combinations)

Identifying the phone and manufacturer (you may need to integrate data from the Wireless Universal Resource File (WURFL)

IP addresses changing based on moving between mobile towers.

Vendors committed to JavaScript page tagging, such asCoremetrics and Omniture, are touting a "no JavaScript" image tag implementation strategy. However, these and other tag-based vendors rely on a JavaScript library file to collect data about the OS, size, and resolution of the monitor. To compensate, you need to stuff the image tag with query strings that will collect the data you require for reporting.

From the perspective of collecting browser (user agent) data, log file and packet sniffing still make sense because these are server-side methods, and therefore the http request gets logged. So AuriQ (packet sniffing), Unica and WebTrends are all possible alternatives, but still suffer the same issues with regard to unique visitor identification.

On the reporting side of things, vendors will point out that you can get the same reports as you'd get normally, which is true; you just have to filter accordingly to get a break out of mobile activity. Nedstat is an exception, recently announcing a more packaged treatment of their mobile analytics reports.

Reflecting the ongoing industry fragmentation within web analytics, a collection of new firms have emerged with a more specialized approach to analyzing the mobile web. I'll take a look at these companies in my next post.

Vendor criticism of CMS Watch

aps@cmswatch.com(Alan Pelz-Sharpe) — Fri, 16 May 2008 10:43:00 -0400

As you know at CMS Watch we write critical product evaluations to help you avoid expensive procurement and deployment mistakes. We write reports that detail both the warts and merits of big vendors like EMC, Oracle, Xerox and IBM -- through to smaller specialist vendors like Hyland, Autonomy and Nuxeo. Readers of our reports often ask me "what did vendor x say when they read that!" The assumption, sometimes correct, is that vendors freak out on reading such criticism.

In an industry whereby most of the "independent analysts" are heavily dependent on revenues from the very firms they claim to be "independent" of, it's unusual to see truly critical research get published. So it becomes a surprise to both buyers and sellers when they read such criticism. In our reports we widely distribute the compliments and brickbats -- if something is truly terrible we will tell you.

But most of the time it is not a case of bad technology versus good technology. Rather it is a case of good fit versus bad fit: a product that could become an outstanding performer in a larger legal firm may make a terrible fit in a mid-sized manufacturing and ERP-centric environment. Hence we urge you the reader to study all the alternatives and balance them out, rather than look at one preferred vendor in isolation.

Speaking of isolation, the marketing groups of some vendors seem to operate in in a kind of vacuum. I guess it's part of the job for them to drink their own Kool Aid, but some of them seem to think it's part of their job to attack and stop any criticism of their product or company. At CMS Watch we're often on the receiving end of that wrath; that stinks sometimes, but so be it. Just as it is the vendor's job to wax lyrical about the joys of their product, so too is it ours to unearth the reality. If you want to get an insight into this particular dynamic, whether you're a curious end user or a vendor AR (Analyst Relations) person, check out the article I published today.

Omniture's Q1 results -- forecasting the future of analytics?

philkemelor@pkwc.com(Phil Kemelor) — Thu, 15 May 2008 23:41:00 -0400

Omniture's Q1 2008 financial results were as rosy as ever, and after listening to a year's worth of earning's calls, the transcripts are almost interchangeable...higher revenues, higher profit margins, more sales reps, more upselling of Genesis partner services.

Omniture's product strategy is focused on providing options and services on integrating web data to both marketing and offline data, the logical next step for web analytics that we discussed in the first edition of the Web Analytics Report.

Yet, how much of the potential marketplace for analytics is actually going down the integration path? Is your organization doing this? Or, does integration simply mean linking web analytics with Google AdWords?

Omniture is a bellwether company in the web analytics industry, but one trying to execute on the promise of web analytics integration in order to sustain its growth. Other vendors are following the same strategy. This could well become the final frontier for traditional web analytics vendors.

In a market that is now 13 years old, web analytics is no longer in its infancy. It may in fact be plateauing in features that are of real importance. Now we'll see if customers can finally catch up and use the data to drive web value.

SEO: furthering the case for better content hygiene

tregli@cmswatch.com(Theresa Regli) — Fri, 9 May 2008 16:12:00 -0400

The worlds of SEO and enterprise search are not as far apart as you might think. Let me explain. Along with my colleague Alan, I had the pleasure of attending the upbeat Internet World UK conference earlier this month in London. What I enjoy most about attending such large, diverse events is going to sessions about technologies related to, but not directly about, the technologies I cover -- and so I found myself attending numerous sessions on SEO, or Search Engine Optimization.

Though there was an inquisitive crowd present for my talk, the topic I spoke about -- enterprise search -- is much lower on the totem pole for e-marketers (the principal attendees at this event) than SEO. A few showed up, in fact, wanting to better understand the nuances of website search vs. enterprise search vs. SEO. Are there universal things we can do to simultaneously improve all things, I was asked? Indeed. Despite the differences in searching the web, searching within your enterprise, and making your own website more findable by search engines, many of the same best practices apply.

As readers of The Enterprise Search Report know, good content hygiene is essential to good search results. Consistent content structure, metadata, and simple things like clear and meaningful document titles all help search engines work better -- be they the public (e.g., Google) website, or enterprise kind. But in most cases, content managers don't know where to start with their clean-up: it's an often overwhelming task. As such, content clean-up continues to get cast aside ("too much work"), and search technology vendors make it seem that much less important when they promote technology as the panacea to content woes. Don't believe it for a second.

A few more tidbits from the event:

These days, 80% of e-commerce transactions start with a web search, says Dan Cohen, Head of SEO for MSN UK.
The #1 hindrance to SEO is poor content and code (which also highly contributes to poor web and intranet search results)
Creating clear, topic-related content "hubs" on your web site is the next most important thing for SEO
As readers of our Web Analytics Report know, analytics should be a core part of your e-marketing pie, to get a clear picture of what your customers are looking for and where they're getting stuck.

Bottom line: cleaning up your content can improve your website search results and your Google ranking.

Coremetrics releases ad hoc analysis functionality

philkemelor@pkwc.com(Phil Kemelor) — Thu, 8 May 2008 06:14:00 -0400

The recently concluded eMetrics Summit was somewhat quiet on the vendor front except for the Coremetrics announcement of its Spring 2008 release featuring ad hoc query functionality in its new offering, "Explore."

This now gives Coremetrics a competitive offering to Omniture Discover and WebTrends Market Intelligence.

More advanced users of Coremetrics will likely be pleased. As Web Analytics Report readers know, one of the major pain points among customers seeking deeper-dive analysis was needing to go through account managers to run custom queries and potentially have to wait a long time to get results. If Explore works as billed, it will mitigate that problem.

I've always thought that Coremetrics did a decent job at creating out-of-the-box reports that address the needs of marketers and less sophisticated analysts. While they have lagged their competitors in offering a strong ad-hoc analysis tool, another way of looking at it could be that they have been tracking their customers learning curve more accurately...and now they believe there is enough analytics maturity among their client base to actually use a deeper analytics tool.

If this is the case, it represents a different approach than what has often characterized web analytics..."a build it and they will be sold" strategy that provides tools that are too sophisticated and challenging for the customer base to use effectively. As you review web analytics software -- as you would any technology -- make sure not to over-buy something you don't know how to use.

Y! IndexTools...let the games begin

philkemelor@pkwc.com(Phil Kemelor) — Sun, 20 Apr 2008 23:19:00 -0400

The announcement to make Yahoo! IndexTools a free service, coming so quickly on the heels of the acquisition, would seem to serve notice to Omniture, Google Analytics, and other market leaders, such as Nedstat, Coremetrics', Unica, and WebTrends' about the seriousness of Yahoo!'s intentions.

Eric Peterson has written a very thoughtful post that delves into the market implications on this latest move.

But from a customer perspective, this could become a bit confusing in the short term. According to IndexTool's Dennis Mortensen, current customers were contacted to let them know they'd be able to continue using the service at no cost if they sign forthcoming agreement from Yahoo!. Details about the agreement and how this impacts current customization projects is still being sorted out, as is how long customers will have to determine if they want to accept the terms.

I'm sure most customers will be inclined to sign the agreement to maintain continuity, unless they have concerns about Yahoo! storing their data, as Google stores Google Analytics data.

Is this a founded concern?

It all depends on your privacy policies -- something you should consider in your requirements for a web analytics tool to begin with. If you haven't figured this out, then you should.

My guess is that for most current IndexTools customers, this will not be a show stopper.

If it does present difficulties for your enterprise, now would be the time to review the vendor profiles in the Web Analytics Report.

In the meantime, current IndexTools customers will surely be asking some important questions, like whether all their current functionality will remain available for free, and if so, for how long. Will data from the pre-Yahoo! days still be available? For how long? How will this affect custom work that you're doing or planning have done by IndexTools, as well as whether there will be a new technical and professional service availability and cost structure? And what about new features and releases, such as Rubix; what will be the cost and support structure?

Yahoo! is moving quickly, and I expect that they will seek to address these issues. However, as a customer, you'll have to make sure that you get these and other questions answered completely before signing on the dotted line.

Love Your Local Data Warehouse Manager

tbyrne@cmswatch.com(Tony Byrne) — Thu, 17 Apr 2008 11:35:00 -0400

Within most enterprises, the worlds of "data" and "content" remain far apart, despite recurring business needs to converge different types of information. Tony Byrne argues that data specialists have a lot to teach content specialists -- and vice-versa. The trick is finding common ground, allied interests, and a common vocabulary...

Yahoo! steps into analytics with IndexTools acquisition

philkemelor@pkwc.com(Phil Kemelor) — Sun, 13 Apr 2008 10:06:00 -0400

This past week's announcement that Yahoo! purchased IndexTools puts a new spotlight on the web analytics marketplace.

Yahoo! is clearly looking to compete with Google, but the reasons for this particular acquisition remain less evident. On the one hand, Yahoo! may be assuming that the mass market wants the kind of richer features that IndexTools offers. As you raise your own level of analytics competence, you may prove them right. On the other hand, since there were only a handful of independent, mid-range analytics vendors available out there for a decent price, IndexTools may have come to Yahoo! via more of a process of elimination.

The Yahoo! "party line" is that the technology will be a great boon to its small and mid-sized business (SMB) clientele. Probably true. In many ways, IndexTools resembles Google Analytics in its usable interface, featuring both dynamic drilldown and behavioral segmentation, as well as a nice collection of out-of-the-box reports oriented towards campaign analysis.

However, it is the perceived potential of IndexTools that has many observers hoping for more than just another Google Analytics. The company has been touting its next generation release, called "Rubix," since January. If Rubix lives up to its promise, it could possibly give Omniture's Discover offering a run in terms of functionality and ease of use. This has become the second-most anticipated non-release of a product in web analytics -- after Microsoft's Gatineau.

While Rubix could be a differentiator, without it, IndexTools does not offer the functionality that distinguishes it from Omniture and WebTrends -- for example the ability to analyze unaggregated data from a graphic UI and to perform repeatable Excel reporting. For now, you must use regular expressions to analyze unaggregated data and do manual updates of Excel...just like Google Analytics.

Dennis Mortenson, COO of IndexTools, claimed repeatedly that IndexTools could do 80 percent of what Omniture could do, at a fraction of the price. People also say the same about Google Analytics. This is marketing spin at its best. It doesn't matter which 80 percent or which 20 percent; it matters only how it matches your requirements.

As Web Analytics Report readers know, larger IndexTools customers picked that solution to get good standard reports, plus additional reports customized by the vendor, all at an attractive price. Feature richness and attention to individual customer service are not traditionally the hallmark of mass-market solutions, so Yahoo! has some clear choices ahead here, and IndexTools customers will want to watch carefully which way the new owner takes the service.

Among the questions that remain to be answered:. Will Rubix ever see the light of day? Will the basic technology be morphed to a Google Analytics-type solution? A combination of the two perhaps? Or will all of this become moot if Microsoft acquires Yahoo!? Or perhaps IndexTools becomes the premier analytics offering from Microsoft?

We'll be watching.

Readers' challenge - name our new chart!

aps@cmswatch.com(Alan Pelz-Sharpe) — Sun, 6 Apr 2008 17:11:00 -0400

Since moving to the US in 2002 I have become a fan of The Atlantic magazine, and in particular the last page of each edition that contains the "Word Fugitives" column. In this column readers ask for new words to meet commonly demanded needs.

For example, in the most recent issue somebody requested a polite but meaningful phrase for a couple to use when they are trying to conceive a child. My favorite of the suggested responses came from Laura Whitman, of Redwood City, Calif., who wrote, "In our group when a couple is married and everyone is wondering what their plans are in regard to procreation, we always ask if they are in the product research or product development phase...early product development refers to your correspondent's criteria." You see there is always somebody out there with the right answer.

So to our current challenge, we have designed a wonderful (in our own opinion) chart that provides buyers of technology with an at-a-glance risk/opportunity review of the products we cover in a particular segment. We refer to this currently as the Vendor Risk Report -- and internally as the "VORP" (Vendor Opportunity and Risk Profile). Not exactly terms that roll off the tongue. Now some of you may be aware of similar chart like products from other firms, charts that have catchy names like "Magic Quadrant" or "Wave." Well we want a catchy name too but haven't thought of a satisfactory one.

So in the spirit of Word Fugitives, we are throwing the door open to you. Please send me your thoughts and ideas and we promise to publish and credit the best of them here on the site. And for the very best one we will send a box of gourmet chocolates or bottle of good champagne (your choice). The gauntlet has been thrown down, can you rise to the challenge?

IndexTools and WAA Standards

philkemelor@pkwc.com(Phil Kemelor) — Tue, 25 Mar 2008 23:21:00 -0400

Given pervasive confusion around analytics terminology, I lauded the Web Analytics Association's August, 2007 announcement of Report Definition Standards, but was somewhat skeptical with regard to how vendors might use the cloak of compliance to make it harder for you to interpret their report definitions.

Fast forward to March 9, 2008, and IndexTools' COO Dennis Mortensen posted a refreshingly comprehensive, easy to understand, and transparent list of how the IndexTool's solution complies (or doesn't) with the standards.

Critically, Mortensen describes the methodology IndexTools uses to make the calculations, which lies at the crux of using the standards definitions to understand reports generated from any product.

Mortensen's post must also rank as an industry first for a vendor to actually go public that their product cannot produce every metric under the sun, making it far easier for you to potentially draw an "apples to apples" comparison and clearly see what you get, and what you don't, from a basic reporting perspective.

All of the other web analytics vendors...especially those with the largest customer bases, such as Omniture, Google Analytics, Nedstat, ClickTracks, Unica, Coremetrics' and WebTrends should step up and clarify in similar terms how they meet (or not) these standards.

Nobody's really number 1 in Web Analytics

philkemelor@pkwc.com(Phil Kemelor) — Fri, 21 Mar 2008 00:27:00 -0400

"Who's number 1?"

It's question people ask a lot in many domains, but especially software, and as such, it regularly pops up within the web analytics community.

You can rank them crudely by number of individual customers. Let's take a look at the vendors we reviewed in the most recent Web Analytics Report and see how many customers they have -- or rather, they say they have:

Google Analytics: ~~5 million~~ "hundreds of thousands" (Enterprise/SMB)
WebTrends: 10,000 (Enterprise/SMB)
ClickTracks: 8,000 (SMB)
Nedstat: 1,200 (Enterprise) 6,000 (SMB)
Omniture: 4,000 plus(includes Visual Sciences) (Enterprise/SMB)
IndexTools: 3,000 (Enterprise/SMB)
Unica: 500 (Enterprise), few thousand (SMB)
Coremetrics: 1,200 (Enterprise/SMB)
Fireclick, : 230 (Enterprise/SMB)
AuriQ: 200 (SMB/Enterprise)
Foviance: 50 (Enterprise)

It's an interesting list, but what does it really tell you? It tells me that some products target different sizes of customers and may be more niche than others. But can you really say who's number 1 or who's winning? I'd say these results only make for good conversation, just as picking who'll win the World Cup or this or that election.

Would you buy a solution because it's perceived to be the market leader? Because the buzz is that "everyone is buying" that solution? You shouldn't. I find that vendor selection is often given short shrift in today's market because of a perception that there are only a few real choices.That's simply not true. You have some very important choices.

Figure out your requirements and manage your vendor selection process. Don't simply favor one vendor over another because it's getting all of the good press clippings, and nods from the Wall St. analysts. In the long run, you'll be much happier with your analytics tool selection. Or, to put it another way, the "Number 1" vendor is always whichever one that's a right fit for you.

PS: Have you signed up for my web analytics class in Copenhagen?

Web Analytics Class in Copenhagen

philkemelor@pkwc.com(Phil Kemelor) — Fri, 14 Mar 2008 00:57:00 -0400

There's a lot more to web analytics than simply running software. So please join me at a special full day web analytics workshop in Copenhagen on April 16th.

Improving Web Site ROI with Web Analytics will be an intensive treatment on what you have to do to:

establish and maintain a successful web analytics program
develop key metrics and reports
apply analysis to improve website effectiveness, customer satisfaction, and online marketing

Of course there's more. To find out all the details and register, go to:the class information and registration page.

I'll be sharing best practices, tips, and techniques for to apply as soon as you get back to your office.

Omniture's SiteCatalyst 14 Catches Up

philkemelor@pkwc.com(Phil Kemelor) — Thu, 6 Mar 2008 00:07:00 -0500

Everybody loves a party, and Web Analytics vendor Omniture is no exception, using their annual summit to announce the release of SiteCatalyst 14 and a truckload of other news, including a partnership with Baidu and the roll out of the combined Touchclarity/Offermatica offering.

If you have SiteCatalyst, you probably are now just getting familiar with the new release. So, what do you think of it?

I received a briefing last week, and certainly the new Ajax interface jumped out as a big improvement...13.5 was getting tiresome to look at, especially when compared to Google Analytics, IndexTools, and Nedstat.

SiteCatalyst's new capabilities in video measurement are also worth delving into. Tagging video is generally labor intensive because tags must be set within the application. Omniture claims to enable a "one tag" fits-all approach for Flash and Flex, along with Windows Media Player, QuickTime, and RealPlayer, which appears to be similar to the approach taken by Google Analytics. And there are improvements to Excel reporting and Forum access that are also noteworthy. To see some video of the new reporting, check out Marshall Sponder's blog.

If you're an Omniture customer, you'd have to be pleased with this upgrade. If you're an HBX customer contemplating whether to stick with the plan to migrate to SiteCatalyst, this will also appear to be fairly compelling from a features perspective at first glance.

Nevertheless, I suggest you take a closer look at what's really important from a requirements perspective...Is video analytics really something critical for your business? Do you use Excel reporting? Could Google Analytics fulfill 90 percent of your requirements and then a video analytics company like Visible Measures fill the last 10 percent? These type of mix and match possibilities can make your head hurt, but finding the right combination of measurement tools is going to get more, not less challenging, in the years ahead.