There are certain actions that need to happen before a visitor can purchase—visit a product detail page, add to cart, checkout, make it through checkout and then purchase. To decide on how to optimize an ecommerce website, it’s helpful to narrow down the options of what to work on first by focusing on these required steps. Every section of the site then has a purpose: getting the user closer to taking one of those necessary steps.

When analysing a homepage or category grid page, focus on making changes that will get more visits to the product detail page.
When analysing a product detail page, focus on making changes that will increase the amount of visits that will click the add to cart button.
On the cart – what’s going to get more clicks on checkout?
In check out – what’s going to get more clicks on the order submit button?

There is plenty to do to decrease friction and increase motivation on each site section to get visitors from one step to the next. The best way to prioritize is to work backwards. When a visitor first gets to the site, their purpose is very unclear, there are a million different reasons why that visitor showed up—how do you optimize for that? By comparison, when a user is in the checkout process, their motivation is high and their purpose is clear – they want to checkout with the items they’ve selected. So the ideas of what to do to increase checkout conversion are easier to get to: improve the form fields, decrease distraction, increase confidence.

Expectations for web analytics is often misplaced. With so much data available, the answer to the question, “what is causing the site to make/lose money” must be in there somewhere. The truth is that the answer is not in the data to begin with. That’s not the point of the tool. It doesn’t give you answers, it gives you questions to ask.

The data will allow you to form a hypothesis that you can use to test changes on the site, and then measure what happened. “I wonder if we can attract even more traffic if we create more content like the kind that has the highest visits.” “I wonder if making the add to cart button a more prominent color will increase products added to cart to reverse this negative trend in conversion rate.” “I wonder if making the guest checkout button appear first will decrease the amount of shoppers that abandon the first step of the checkout process.”

Data isn’t the solution. It’s the ideas that come from the data, applied to see what happens next, that is the solution. Small data can instigate just as much of a good idea for what to do next as big data.

Learning from your customers to improve your website can be gathered in multiple ways. Each of the four quadrants in the diagram below provides different kinds of information, each important in it’s own way.

In the Attitudinal half, you receive information about what people say they would do or what they say they want. A lot of surveys and interviews can be unreliable. User’s memories are fallible and can miss details that are important to design. People also say things that they think you want to hear so as to not be impolite or against the grain. That said, some simple survey questions can be revelatory when related to user’s motivation, as in “why are you visiting the site today?” This combined with another simple question, “were you able to accomplish your task?” can give great insight.

The behavioral half gives you the advantage of seeing what users actually do not just what they say they’ll do. Web analytics can give huge amounts of quantitative data that you can’t find anywhere else, but the caveat with all that data is that it doesn’t tell you why like surveys will and it doesn’t tell you how. Deciphering how users interact with a product detail page or how they navigate from page to page isn’t easy with web analytics. There are pathing reports but they don’t help much. This is where usability testing comes in. It allows you to watch users and learn how they attempt to perform tasks. This gives you insights that aren’t accessible from any other source.

Relying solely on one source would be a mistake. Combining any two is worth more than the sum of their parts.

Heretofore is my Law of Diminishing Marginal Utility Of Web Analytics Reporting: The larger the distribution list of your web analytics report, the less valuable that report is to any individual.

Wherever there is web analytics there are big, watered-down reports being generated and sent to lots of people every monday – it doesn’t have to be this way. Sending out reports to the entire company starts with the best intentions. Everyone wants to know what is going on with the website. Requests are made to include a little bit of everything for everyone: overview of total traffic and conversion rate, ROAS for the marketing team, best selling products for merchandising, cart conversion funnel for the UX team, social likes and tweet counts for the brand team, site response times for web development, customer service stats, and the list goes on. Pretty soon it’s a report that takes hours to create and when it arrives Monday morning it is swiftly ignored by everyone in the company.

One of the basic productivity concepts of David Allen’s book Getting Things Done, is to spread your actions over all the contexts you are acting in. Your daily actions happen in some very clearly separated areas, called contexts: @Home, @Work, @Computer @Store, for example. It doesn’t do you any good to remember to buy toilet paper when you’re at work. So on your to-do list you can tag your action to buy toilet paper with @store, and when you are in that specific context you only need to focus on those applicable actions. The same concept can be applied to web analytics reporting.

The right data needs to be available at the right time to the right person. If any of those three don’t match up, the report is just noise.

1. Only data that is actionable on behalf of the person consuming it is needed. Consider only consuming information that applies to your circle of influence. The point of web analytics data is to incite change. If you do not have access to levers then what good does the data do you if you can’t pull them? Maybe someday you will have access to the levers and the data will wait happily for you until then.
2. Monday is an arbitrary day to look at your reports unless you only make decisions on what to do next on Mondays. Your job probably requires looking at the data that matters to you at different times, all the time.
3. The right data is discovered when you have the right person in the room and you’re able to discuss what business objectives they have that can be mapped to trackable site behaviors and what levers they are in charge of.

Looking at the criteria of A, B and C, a weekly mega-report is pointless. Full of nothing helpful to anyone. instead, individualized reports tailored to the user, available and accessible when the user needs them, is the only rational approach.

Web analytics tools suffer from a lack of constraint. With so many things that can be tracked, all of it is and served up all at once to the unsuspecting user who opens up the reporting interface. The amount of data creates a nagging feeling that will start to follow you around, telling you that you’re not using the tool to it’s full potential. Pretty soon you start coming up with business objectives to match the reports you have instead of focusing on your business objectives and then reporting on them second. Eventually you’re looking at everything and not deciding on anything.

Just because there are top entry pages, cart additions, time spent per visit, device types, paths, internal search terms, referring URLs and on and on, it doesn’t mean you need to pay attention to them. At least not right now, and especially not all at once.

From Seth Godin:
“If you build your company with the policy that you’ll never run an ad, it makes it even more important that you build a remarkable product–you’ll never be tempted to compromise and try to make it up with hype. Same thing goes for organizations that refuse to pay bribes. By eliminating situational decisions and grey areas, it changes strategies from the top down. Or perhaps you’re not willing to pay overtime, regardless of the emergency, regardless of how late the project is… it makes it far more likely projects won’t be late, because they’re designed to ship without emergency… Rigidity is rarely your friend, but well understood boundaries make decision making a lot easier.”

Focus on one site section or source of traffic along with a couple metrics. Make changes to that site section or inbound traffic source. Measure your success and then move to the next one. All those other metrics and reports in the interface will happily continue to gather data and wait for you to make them relevant.

With so much data available with web analytics tools, one would think that businesses would become more confident in their understanding of themselves and their customers. But more data leads to just as much uncertainty which is described with a perfect analogy in this New York Times article.

Here’s the analogy:

“The larger the island of knowledge grows, the longer the shoreline — where knowledge meets ignorance — extends. The more we know, the more we can ask. Questions don’t give way to answers so much as the two proliferate together. Answers breed questions. Curiosity isn’t merely a static disposition but rather a passion of the mind that is ceaselessly earned and nurtured.

Mapping the coast of the island of knowledge, to continue the metaphor, requires a grasp of the psychology of ambiguity. The ever-expanding shoreline, where questions are born of answers, is terrain characterized by vague and conflicting information. The resulting state of uncertainty, psychologists have shown, intensifies our emotions: not only exhilaration and surprise, but also confusion and frustration.”

Exhilaration and surprise on one side, confusion and frustration on the other – as any web analyst would attest – sums up web analytics perfectly.