10 Most Common Web Analytics Mistakes


So you want to know the most frequent questions I get about web analytics and online marketing? I’ve tried to present them here and answer a few of them so maybe I can save you, my clients and future prospects some time so we can get onto executing profitable campaigns.

1) Free analytics are good for my business and are just as good as paid ones

Coremetrics Platform offers more than Google Analytics

First, let me say, I like Google Analytics and Yahoo’s Analytics solutions. But Google Analytics was built for Google. It serves their interest, not yours.

This might be the most common myth out there, and possibly one of the most dangerous. First and foremost, to directly dispel this rumor, the truth is that enterprise analytics solutions offer levels of detail and capabilities that free solutions such as Google Analytics don’t. Although free solutions can be hacked to a certain extent to extend their capabilities, at the end of the day they fall far short of the paid solutions in providing meaningful reports to a diversified group of stakeholders. If for example you had any interest in tracking the video drop off rates, or immediately integrating custom analytics events, records or values with a recommendations engines, you can’t.

The Omniture Online Marketing Suite offers a platform to manage the online marketing discipline for sophisticated digital marketers.

Beyond capabilities alone, you also get a lot of important side benefits as well. First amongst them is accountability and support. With free solutions you’re not paying for anything, so you get a community of blogs and consultants, which we belong to actually. If you’re paying for a solution, and if for any reason it has a problem, we’ve got help from the analytics provider to get the job done.

With paid solutions you own your data and can integrate it seamlessly with PPC bid management platforms, multi-variate testing platforms, and a host of other add-ons and integrations like Salesforce, NetSuite, SugarCRM, email marketing solutions, ad-server networks and more.

The future of analytics is in an analytics suite approach, not in point solutions. Of course, if you want to own your own data from Google Analytics, they offer a license to purchase Urchin for approximately $3,400 (the company that Google acquired in 2004 that is today known as Google Analytics). The future of analytics isn’t more data, but instead the ability to act upon that data in an a fast, efficient, and measurable way. Free solutions will never offer the level of integrated optimization that paid solutions do, and due to the great competitive advantage that such features offer, a business should be wary of the risk that going “free” might entail. Bottom line: If your business relies on the internet, you should investigate using a paid analytics platform. I have seen cases where the answer is to stay on a free solution, but more often than not, it’s time to increase conversions and there’s too much at stake to penny pinch.

2) There are a couple of KPIs that are silver bullets

About the only silver bullet in marketing I can think of is the word “FREE”. That’s as close as it gets. The “best KPI’s” are the one that works for your business. No “standard set” of KPIs exists for all business. However, we do believe that there are best practices by business model. E-commerce companies have a clear set that are extensions of the purchasing funnel. Professional services and lead generation companies have a similar but varied set of web analytics KPIs and of course, companies with business models in advertising and the distribution of paid content have an engagement funnel filled with KPIs. Many marketing managers seize hold of a few KPIs without regard to their business model and lose sight of what is truly important for their business.

Sometime they’re lucky, and those metrics are indeed what they needed. Every business has variations in their purchasing funnel or engagement funnel. For example, a sheet music company may have a “recommendations engine” that drives a lot of their business while a fence builder may have a “configure your own fence tool” as a key driver towards conversion. These individual features are what make believing in blanket “silver bullet KPIs” wrong.

When defining the KPIs important to a business, its wise to look at examples, but one must always remember that each business is unique and what will for work it might not be what worked for others. Always approach your analytics implementations with that in mind, because no matter how good the tool might be, if you don’t know how to use it for your project, you won’t get the value you seek. This leads us directly into our next myth:

3) Web analytics tools are overwhelming because there is no way to make sense of all the data they collect.

If this were a movie, picture me saying “Noooooo” in slow motion while I throw a lifesaver to my friends drowning in data. Don’t give up and don’t try to eat the elephant in one bite. Every website can generate a mountain of data. Most of it isn’t important. We once had a client ask about the speed of mobile browsers and whether they wanted to know about the wireless protocol used by each carrier and what average speeds were for those phones and carriers. We could measure it no doubt and with Omniture it is an out of the box report, but unless you make a lot of videos (which this client did), it’s a red herring.

Good analytics implementations empower marketers to take action. The outcomes of an analytics implementation are easy, they are to produce the reports, dashboards and alerts that enable marketers to take actions that improve customer experiences and improve the site objectives.

Because you want the nuggets, not the haystack, its extra important that you take the time to understand your business’ needs before you begin an analytics implementation. Taking extra hours at the front-end of the process, will save you more money and effort than anything else at the end.

If you can define the KPIs your business really needs upfront, you’ll avoid getting drowned in unnecessary data. Set up analytics from the get-go to give you the information you really need and you’ll be able to get a greater and faster return on your investment.

4) If we send our tech resources to analytics training we can make more money

That is half of the story. Having a thorough technical implementation is undoubtedly important. But you need a business analyst as well to define what data you should even be gathering, and then analyze and sift through that information to ask the right business questions and then find the areas for action.

It is also important to benchmark and compare. As we said before, the mountain of data is useless without distilling it into actionable intelligence. You need resources that know how to look at your data, compare segments, compare timeframes, products, and so forth.

Analytics isn’t just about data, comprehensive data analysis is just as important, or possibly even more important than getting the data itself.

We use a model that takes data, turns it into information, and then positions your marketers to take action and perform advisory services within the organization as the voice of the customer.

5) We need to choose an analytics solution that has a benchmark capability with other companies in our industry

A few solutions like Google Analytics and IBM’s Coremetrics have insightful benchmarking capabilities. The one from Google offers little in the way of value because websites are grouped according to how Google sees you, not according to how a business wishes to be classified. However, the Coremetrics benchmarking feature is a bit different and quite insightful for those companies wishing to participate in the network. Only companies that are willing to participate in the benchmarking program are allowed to see the benchmarks for that industry and sector. So it’s a nice value-add offered from Coremetrics but shouldn’t be the primary reason for selecting an analytics vendor.

6) Its going to be easy to take action on what we learn from our analytics

Analytics are the tip of the iceberg. Many organizations enter an analytics implementation expecting to get some valuable, but relatively easy/cheap actionable intelligence. And there are indeed a lot of quick wins to be had in most cases.

You’re always going to be getting small suggestions from your analytics like that. However, many times, especially when implementing a paid analytics tool for the first time, an organization should be prepared to take action on more challenging, disruptive suggestions that they might not have been expecting.

Sometimes its not a question of just updating your shopping cart, sometimes the data might be telling you that you need to completely restructure the entire organizing principles of your website. If your company is in its first generation of analytics, it’s likely that you are seeing for the first time how consumers are responding to your website. In many cases, companies built their website in terms of how they view their company, not how consumers see the problem they want to solve. A site redesign is frequently an action that comes based on the data after a solid implementation is performed. Although there isn’t much that can be done to prepare for such an eventuality, an organization should at least keep in mind (and hopefully in their budget) that their newfound insight might not be easy/cheap to act upon. But if they really want to get the ROI they wanted, they might have to commit invest in significantly more than just the analytics itself.

7) Dashboards aren’t for management

Dashboards are fantastic tools for keeping many levels of the organization, from operational, tactical and strategic resources, informed of key metrics in a timely manner. However dashboards should be customized for these roles. Each stakeholder in an organization doesn’t have to see all the information available. One executive might need KPIs such as Revenues, Orders and Units sold and Return on Ad Spend by channel (PPC, SEO, email, etc…) A mid-level manager may want to track this detail by campaign and a product manager will want these same views but filtered by segments. A regional marketing manager needs yet an additional filter for the geography. Tailor dashboards to give those people the information they need to be successful in their tasks.

8 ) Everyone is going to have to log into the analytics every day

We believe that successful analytics implementations should minimize the need for all but the most analytical resources to ever login to the platform. If we’ve done our job right, each stakeholder should have a set of reports and/or a dashboard delivered to them at the right frequency in the format of their choosing.

So dashboards are for everyone. They just need to be customized. If done correctly, they can be some of the most valuable tools analytics offer and while keeping everyone in the organization up as the speed of business accelerates.

9) Its really important to measure unique visitors

Just like we detailed in Myth #2, there are no silver bullet KPIs. This is one of the most common ones we run into. Mostly because these kinds of KPI assumptions come from what we call “Caveman Analytics.” Back in the day, not too long ago actually, analytics solutions couldn’t offer much more than the basic data like Unique Visitors, pageviews, and other such basic metrics.

Look past those old school KPIs, and drill down towards what really matters: what is bringing in revenue. For some, like ad supported websites, Unique Visitors or pageviews are exactly that, but for many its not. If you’re in eCommerce that’s easy to see. In cases like that you should be looking at units ordered, average order values, cart abandonments, cross-sells, and other such metrics. Develop a funnel using our purchasing funnel methodology and determine the cost to acquire a visitor at each step in your funnel.

10) The best KPIs are single data points (ex. gross revenues)

Although data points can undoubtedly be useful, its more important to step back and try to look at ratios and averages. This isn’t only due to the fact that single data points are many times unrepresentative, but also because ratios and averages are much easier to take action upon.

This is because they are more specific. If you ask yourself “How do I raise gross revenues?” you’re going to have a hard time getting an answer. It’s too big of a question because its too big of single data point. However if instead you ask yourself “How do I raise average order values?” you’re in a much better position to take positive action. With a much more specific question you can quickly start try out solutions like recommendations engines, variable pricing, promotions, etc.

Bonus Tip: If I get good at analytics, Executives will label me as being too tactical and not strategic.

This actually happens a lot to folks who are willing to roll-up their sleeves and dive into the numbers.  Make sure that when web analytics data is presented in meetings, that you start off with a very simple fact and then substantiate it.  How analysts communicate with management has a lot to do with how they are perceived.  When those same executives really needed advice on what to do, you’ll be on their short-list of trusted advisors.


Organizations that master analytics do it as part of their routine processes.  Taking action on web analytics data is not a project, but an on-going process.

Posted in

Leave a Comment

Your email address will not be published. Required fields are marked *