The sky is falling! Six hundred thousand workers lost their jobs in January alone – with seemingly no end in sight. We appear to be in the middle of the worst downturn since the Great Depression. Awful news for businesses, right? Well, not necessarily for the analytics business!

A few weeks ago, I read an interesting prediction by IDC which stated that companies that significantly reduce their sales and marketing investment in 2009 will be gone by 2010!

Given we all seem to have tighter marketing budgets these days, what do we do to not only stay in the game but try to gain some ground? This is where analytics comes in: over the past decade, analytics has transformed “the art of marketing” into “the science of marketing”…and when the economy gets as uncertain as it is today, any marketing effort supported by quantitative data (read analytics) is bound to win more attention from those holding the purse strings.

The great strength of analytics is that it demands accountability. A scientific approach to marketing–as opposed to a “gee, ain’t it pretty?” approach–enables continuous optimization in the sales (or conversion) cycle. It dramatically increases marketing productivity by improving both the volume and quality of the leads/sales we generate.

In a culture led by analytics, decisions are made based on facts, not by the loudest voice or the owner of the biggest title. Sentiment gives way to rigorous testing. Analytics-driven companies strive to find out what’s most effective in converting their intended audience. And thanks to analytics, they can track which internet properties or campaigns are producing the quality leads they promised–and which ones are simply NOT delivering. That’s called accountability!

Like many of you, I’m an experienced digital marketer with a budget to answer for. Armed with analytics results, I approach meetings with my media reps confidently. With data backing me up, I know I am spending my precious marketing dollars wisely: renewing only what’s working and renegotiating rates or canceling contracts on what’s not. What’s even more satisfying, I get to show the CEO exactly how our marketing dollars are contributing to overall revenue generation–and even justifying more spend!

So…yes, certainity about marketing effectiveness–which can only be achieved through analytics–is definitely something to invest in, especially in these uncertain times.

There is a scene from Lewis Carroll’s Alice’s Adventures in Wonderland between Alice and the Cheshire Cat that goes something like this:

“Would you tell me, please, which way I ought to go from here?” [said Alice].

“That depends a good deal on where you want to get to” said the Cat.

“I don’t much care where “ said Alice.

“Then it doesn’t matter which way you go” said the Cat.

And if you’re embarking on an analytics project, I would make the same point: have a clear goal in mind. Without specific goals, your team will get lost in the minutiae. Without targets, there’s no telling where you will end up. As the rest of Alice’s conversation goes:

” so long as I get somewhere” [she said].

“Oh, you’re sure to do that,” said the Cat, “if you only walk long enough.”

Unfortunately, “somewhere” may not be where you want your website to be. And sooner would be more desirable than later. To get to a better place more quickly, you will have to set some fairly specific goals. It sounds straightforward enough-but so many of our customers struggle with it, that I thought I’d give some examples.

A simple example of a website improvement goal would be “We want to increase the number of people that purchase our widget from our website by 5% over the next 3 months”.

Here are three reasons why this is a well-stated goal:

1. It is targeted: “We will increase the number of people that purchase our widget from our website”.
2. It contains a specific, attainable number: “by 5%”
3. It has a time-frame: “over the next 3 months”.

Once you’ve set the goal, it is equally important to focus reporting on your goal. Make sure you set up and review reports that are relevant to your goals. Most tools can provide lots of data-you’ll need to filter most of that out and not get distracted by it. And if your tool has the ability to do custom reports, these can be extremely useful to focus your metrics on achieving specific goals.

I’ve seen quite a few people get a new analytics tool, then go through their reports looking for things in their business that will fit the reports. That’s kind of like (in Lewis Carroll’s “Through the Looking Glass”) having the conviction first, then the trial, then the crime.

Analytics tools today offer a Wonderland of data, and it’s easy to get lost and confused. But you don’t have to stay lost.

Set incremental goals. Make sure they are specific, attainable and have a realistic time frame. Once you have defined your goals–then and only then–should you deploy your analytics tools.

Setting realistic goals-or “Key Performance Indicators” if you will-can be a challenge. Often our team at Technology Leaders helps customers define goals-then we deploy analytics against that. When goals are set properly and analytics are geared to support your business goals, the results can be outstanding.

Recently a customer told me he had an internal customer who wanted to know why page views had become an industry standard measure rather than hits.

When I got over being surprised that anyone would still ask this question, I realized there was a bigger issue in play: that of the weakest link. How many analytics professionals find themselves hamstrung by the non-technorati in the firm, asking what appear to be clueless questions? These are teachable moments. It almost seems like the way forward would be to make sure your weakest link in the analytics chain (often the folks who say they “can’t trust” or “can’t understand” the numbers) are not that weak anymore.

If it seems like it’s beneath you to educate these otherwise knowledgeable professionals about an area the understanding of which you take for granted, think about it again. And think about how great it would be to have them become advocates for analytics rather than skeptics (or anchors).

Here, then, is a back-to-basics, sure-to-come-in-handy, copy-and-paste explanation of what the heck a page view is versus what a hit is, and why the former has become the marketer’s most-relied-upon measure of experience (we’ll get to “server calls” on another post, okay?)

I think you’ll be surprised how many non-analytics folks in your firm will be happy someone politely explained it for them.

–

Page View (WebTrends Glossary Definition):

Generally defined as a request to load a single page of a website. On the web, a page request would result from a web surfer clicking on a link on another page that points to the page in question. See also “hit”.

Hit (WebTrends Glossary Definition):

Any request from a file or a web-server. A single page likely contains multiple hits as multiple image and text files are downloaded from the web-server.

My Commentary:

The reason Page Views have become a standard measure is:

–a Page View is closer to a one-to-one  measure of what a user has seen in his/her browser

–a Hit is a  measure of what is being requested of a web server, not of what is seen by a viewer

In addition to this, there are virtually no web marketing measurement paradigms that rely on hits. Analytics reports today, regardless of the vendor, are–for the most part–built around page views.

Typically, there will be fewer page views per unit of time than hits–sometimes significantly. There is no strict relationship between any particular number of hits and any particular number of page views (let’s hold off on what happens to page views when you move from log-files to page-tagging).

All of the above notwithstanding, analytics tools can report hits–but the usefulness of  “hit” data is usually considered to be low due to its relative lack of correlation to user experience.

–

In conclusion, let’s acknowledge that advanced analytics has already moved beyond page-views (for example, what is a “page-view” on a mobile device, or inside a flash movie?). But this post is not about that. It is about making stronger your weakest link–and hoping they will now let you get back to work figuring out the value of your conversions.

Recently,  Forbes dot com published an article about Omniture’s data delivery problems:

http://www.forbes.com/ebusiness/2009/01/20/omniture-web-analytics-tech-ebiz-cx_ag_0121omniture.html

One of the questions the article seemed to pose was: does SaaS (Software-as-a-Service), as a category, confront a particular latency (in this case data-delivery) challenge?

In simpler terms: if I rent my analytics software, is it more likely I will have a problem getting my data than if I own my analytics software?

On a somewhat related note, it was reported elsewhere that online streaming media of the inauguration was a remarkable flop, to wit: “when it comes to large events with millions of simultaneous viewers, broadcast media has nothing to fear”; apparently, despite CNN’s and Akamai’s best efforts, streaming server-arrays seemed to fall over and die after reaching about a million simultaneous calls.

I am illustrating a point: for SaaS (or “on-demand”) as well as streaming media, it’s all about anticipating bandwidth. If you are offering software-as-a-service and plan well, your SaaS offering will not fall over.

Does that mean analytics users should shy away from outsourcing that part of their IT architecture? In general, no. But it does suggest a reason to analyze why you would outsource in the first place.

In order to judge the value of SaaS for analytics, let’s very briefly state why many companies have chosen it: mainly because they couldn’t get their internal IT departments to implement, configure and/or maintain software in a manner commensurate with need. The very complaints about SaaS failure–failure to deliver data–are at the root of having chosen SaaS in the first place. Again, in simple terms: “When I call IT nobody cares–I’m moving to on-demand”.

When you choose an on-demand analytics offering, you are changing just one thing about your analytics architecture: you’re not managing your servers anymore. All the rest is, more or less, the same–the need to define KPIs, tag, configure, maintain; and ultimately rely on someone, somewhere, to make sure the data is delivered.

So the question of whether you’re more or less likely to be deprived of better analytics by choosing an on-demand solution is probably besides the point. You are still going to need your provider–internal or external–to plan well.

Driving in my car the other day, I noticed the “engine temperature” light on the dashboard light up. Only somewhat alarmed, I pulled over to find the cause. It turned out to be a hose leaking coolant causing the engine to begin overheating. I might have risked terrible damage had I not been warned. Then I thought about why it might be a good idea to get warned about function-levels of other valuable mechanisms that impact my life—like my company’s web site.

Most automobiles today have computer controls over nearly ever system in the vehicle, and they continuously monitor the vital signs of the power plant. The tricky thing these computers must do is determine if the current health of the engine is within bounds, or whether something is making its performance fall outside acceptable limits. These limits fluctuate over time as the engine wears, but certain criteria modeling remain in place. When the computer does determine there is something worth notification, it sends up the red flag – usually a light or alarm on the dashboard – to alert the driver.

Can a web site be monitored for these ever fluctuating changes?

With today’s market and business conditions, there is a very high likelihood that your web site is experiencing some sort of fluctuation around the historical trend, and you may even be faced with a reduced staff to monitor these trends. Just as there is an automated system to monitor the health of your vehicle, why not the same kind of automated system to monitor your web site?

Sure, I have a vested interest in what I am going to tell you next, but here’s the thing: there is such a tool. It’s called Dynamic Alert. I happen to be the product manager for it. And I also happen to think it’s a really valuable tool for anyone who wants to spend less time monitoring the overall health of their web marketing effort.

Here is how Dynamic Alert works: by continuously monitoring your existing web analytics data, Dynamic Alert looks for spikes that lie outside the dynamically generated limits, and creates these limits based on (among other things) historical trends. If an outlier is detected, Dynamic Alert sends an email notification (an “Alert”) with the specific dimension and measure combination deviating from the historical trend and limits. In other words, it tells you something happened and what.

Besides being told that yesterday’s campaign was a success (the increase in revenue measure triggered an Alert!), wouldn’t you also want to know that the changes made to the check out process broke the tagging and will cause errors in reporting? There are so many more applications for Dynamic Alert, it would take more than one blog post just to list them.

Just like the temperature warning light on your dashboard lights up before the engine overheats, wouldn’t you rather be warned of an issue BEFORE irreparable damage is done and someone asks to see the reports?

Dynamic Alert was created by Technology Leaders. You can get it directly from them, or you can also purchase it through WebTrends.

We’ve come across a common occurrence at our clients:  one group has implemented tagging from one vendor and another group overlays the site with tagging from a different vendor.  In theory you could take a single analytics vendor and use them to provide all the web analytic information you need, and most analytics vendors recommend that.   However, we’ve found that there are often different constituents for Web Analytics results who sometimes don’t have the same requirements. For example, a marketing group would be primarily interested in the customer facing data using client-side tagging while an IT web server support group might want to look at reports that show server utilization.  So, the question is whether to use one analytics vendor and tagging strategy for all purposes on a web site or to use multiple methods.

This has often been structured as a question of whether to tag the site (for client interactions) or to analyze the server logs, but with the advent of free client-side tools such as Google Analytics, it is being raised for client-side environments as well: how do you correlate the data from two different analytic tools on the same page. We have come across this situation with many clients and they frequently find that the numbers do not directly correspond.  One of the major reasons for that is because the different analytics vendors sometimes have different criteria to identify visitors:  in a multi-domain organization where visitors move from site to site, strict First Party Cookie identification needs leverage from a Third Party Cookie to track the same individual across multiple sites.

We’ve also found that when tagging on the web site is done by different groups (which is often the case when different analytic tools are used), the tags don’t always match in content or interpretation.  This can become an organizational problem, but once pointed out, it can be usually addressed.  That said, it still is useful to tag the same pages using different tools to

• Crosscheck advanced reporting against basic reporting
• Leverage specific reports from different vendors
• Correlate different data sources for integrated reporting (Geographic data can come from different sources depending on the web analytics vendor.

In the end though it is best to look at complex tagging from only one source and leave any alternate analytics tags at the simple page oriented level to reduce the effort in correctly tagging actions of interest by targeted individuals.  So our position is that while you would not want to have to maintain multiple tagging sources for the complex conversion tracking, it is still useful to have a basic tag from two different sources to cross check the vendors.

Can marketers benefit from another blog on the subject of web analytics?

We say they can.

Technology Leaders (TL) has been deeply involved in analytics since 2002. We’ve seen the industry grow from a small coterie of technical insiders to a mature, multifaceted industry that informs key marketing and business decisions at nearly every important organization in the world.

When we helped found the Web Analytics Association in 2005, there were three of us: Bryan Eisenberg, Jim Sterne and myself. Now there are over 1,500 members worldwide. And the most important industry event, eMetrics, has grown from a comfy retreat at a seaside resort in California, to a worldwide juggernaut with outposts in Washington DC, Toronto, and Europe.

TL’s other two partners, David Millrod and David Smith, have been personally involved in building and extending some of the most robust analytics programs in the world, and offer a truly unique view on what works in analytics, and what does not.

Our board includes two of the most dynamic, groundbreaking individuals in web analytics: Rand Schulman (who encouraged us to start the WAA)–former CMO at both WebTrends and WebSideStory; and Pelin Wood Thorogood, who was Senior Vice President of Marketing at WebSideStory.

All the the folks named above may contribute to TLview from time to time.

TL has recently launched a groundbreaking tool: Dynamic Alert. It helps you find the news inside your web analytics reports, without having to look at all the reports. Recently, WebTrends has agreed to offer it directly to its customer base.

With both services and software in our toolkit, we’ll take a comprehensive view of the world of web analytics. We think marketers everywhere will benefit.