in 25 Books in 2012

#20/25: Web Analytics Action Hero


The original value proposition of web analytics was to help companies achieve their online business goals and maximize the return on their digital marketing investments. Web analytics achieved these objectives by helping companies to measure, analyze, and optimize online performance. In other words, CMOs invested in web analytics to improve their online marketing initiatives through insights found in the online data.

The human touch—creativity, domain knowledge, common sense reasoning—are still needed to realize a greater potential with our data.

Too many companies have shiny, analytics sports cars sitting in the driveways, but no drivers.

You don’t begin your data-driven journey by cherry picking key performance indicators (KPIs) from a list; instead, you start by concentrating on your key marketing goals and objectives

The main goal of any respectable web analyst is to change their organization for the better. How do they do this? They identify strategic, actionable insights and translate them into high-value recommendations, which the business in turn acts on to generate higher revenue, profits, and cost savings.

The difference between a successful and unsuccessful analyst comes down to action that leads to business value. Action is the lynchpin between analysis and value. You may be an excellent analyst, but if your organization takes no action based on your insights and recommendations, you won’t ever be truly successful.

Action heroes are web analysts who are able to get their companies to act on their recommendations and whose analysis work leads to significant returns for their firms. Like the successful actors portraying the popular action heroes, they can command much higher salaries than analysts who may be smarter, better edu- cated, and more experienced but who are not able to translate their abilities into tangible business value.


  • Alignment stage: gather business objectives and create measurement strategy
  • Collection stage: Tagging, data validation and collection
  • Reporting stage: Create and distribute tailored reports

Analysts are caught in a vicious cycle of endless reporting, both maintaining current reports and building new reports. Unfortunately, too many organizations misinterpret reporting as the main destination of implementing an analytics solution and mistakenly view reporting as though it were a part of Actionland. Reports display information, not insights. Information is useful, but insights are invaluable. Insights from analysis can drive actions or optimizations, which can transform a good business into a great one.

Analysis is the isolating of meaningful and actionable insights in data and reports that when acted upon by your organization can drive business value. That’s right: Analysis is the ticket to Actionland.

The deployment of analytics solutions is an ongoing process, not a project you complete once every couple of years.

Whenever sharp analytical minds are spending most or all of their time on report- ing, talent is being wasted. Technology should be doing more of the heavy lifting with reporting through automated reports and dashboards.

Where Do Reporting Robots Come From?

  • Capability mismatch: People don’t have the ability to do WA
  • Wrong structure: owners of WA don’t appreciate value and strategic importance
  • Insufficient training
  • Not enough power: too few people
  • Conflicting priorities: too little time for analysis

The recommendations component is an important differentiator between analysis and reporting as it provides specific guidance on what actions to take based on the key insights found in the data.

In their book Competing on Analytics (Harvard Business School Press, 2007), Thomas Davenport and Jeanne Harris emphasize the importance of trust and credibility between the analyst and decision maker. Decision makers typically don’t have the time or ability to perform analyses themselves.

  • relentlessly and systematically automate as much reporting as possible.

The next step in the two-step rehabilitation process is to inject more insights into your reporting.

Question reporting requirements:

  • What are you trying to achieve?
  • What’s your business question?
  • What actions will you take?
  • How often will you use the data?
  • Who are the key stakeholders that care?
  • Be prepared
  • Ask questions
  • Key pain points?
  • Business challenges?
  • Expected outcome?
  • What is success?
  • Helps to build trust and rapport
  • Ability + Environment + Approach


Business acumen

  • Big picture thinking [Augmentable]
  • Customer-centric [Learnable]
  • Marketing acumen [Learnable]
  • Analytical skills
  • Intelligent [Core]
  • Curious [Core]
  • Attention to detail [Augmentable]
  • Open-minded [Augmentable]
  • Objective [Augmentable]
  • Problem-solving skills [Augmentable]
  • Understand their tools [Learnable]

People skills

  • Communication skills [Augmentable]
  • Interpersonal skills [Augmentable]

Web savviness

  • Online marketing / SEO [Learnable]
  • Online business models [Learnable]
  • Web design [Learnable]
  • Online technology [Learnable]
  • Online trends [Learnable]

HERO factor

  • Passionate [Core]
  • Proactive [Core]
  • Confident [Augmentable]
  • Relentless [Core]
  • Innovative [Core]


  • Corporate goals, e.g. Increase sales, expand product line
  • Department goals, e.g. Increase sales, promote new product line
  • Online goals, e.g. Increase online sales, enhance online customer experience
  • Online initiatives, e.g. Improve site search, reduce shopping cart abandonment
  • Online tactics, e.g. test new landing page concepts, add more editorial content to product pages

By improving your self-awareness, empathy, and social skills, you’ll help clear a path for your high-value analysis.

In order to establish trust-based relationships, you’ll want to gain a deep understanding of their business challenges, manage expectations effectively (under promise, over deliver), keep your commit- ments, communicate regularly, be proactive, and anticipate their business questions and needs.

Your last resort is to pursue a really strong but doomed analysis with the hopes of drawing management attention to the issue. One good opportunity (or more) may need to die so that others can live.

By focusing on your key business objectives and what you’re able to influence, your list of possible targets may be more manageable, but you might still be facing several options.

Starting with the spending side, where has your company made significant investments? The bigger the expenditure, the more attention it should receive.

Individually, the impact of insights gleaned from one particular campaign or microsite might not be a big deal, but if your company is constantly launching new campaigns and microsites then the insights become more valuable.

Companies often neglect older initiatives when internal attention shifts to more exciting, newer projects. […] don’t assume the long-standing areas have been fully optimized.

Conversion funnel:

  • Marketing Campaigns: which are successful?
  • Landing Pages
  • Micro-Conversions: Which conversions are valuable?
  • Checkout/Form Pages:
  • Loyalty Programs

Context matters as it can dramatically impact your credibility, relevance, and efficiency in terms of how you prioritize analysis and invest your time.

  • Change of business goals
  • Ability to Influence – budget surplus
  • Potential Impact – CMO finds out about a new thing
  • Level of Effort – Changes in the tool

Context Wheel

  • Company – strategy, online initiatives, marketing campaigns, website updates, testing efforts, server status, staffing, leadership team, key partnerships, business practices, processes, etc.
  • Industry – Trends in the economy, marketing, consumer, society, industry, etc.
  • Competitors – What are they doing and how are they doing?
  • Tools – How do your tools work and what are their features?

Tribal Knowledge

By plugging into the tribal knowledge (company context), not only can you sharpen your analysis and save time, but you can also better target and prioritize your analysis.

if your analyses seem to be slightly off target and potentially frustrating your audience, verify whether there’s something you just don’t know (maybe they assume you know). You need to shift from “not knowing that you don’t know” to “knowing that you don’t know” and then figuring out what that missing piece of context is.

Try to become the center for online business

  1. Make Observations – are their any trends or irregularities?
  2. Formulate Hypothesis – e.g. broken landing page causes a 80% bounce rate
  3. Gather evidence – e.g. run a test
  4. Draw conclusions

How to avoid human error in analysis?

  • Understand the business challenge / question
  • Do your analysis as uninterrupted as possible
  • If you’re stressed or tired, take a break or continue on the next day
  • Document what you did, so that you can reproduce your analysis
  • Share your analysis with colleagues to catch errors

Sacred cows often make great burgers for analysts and are an excellent source of optimization nutrients. Any time you identify something in the data that contradicts or is inconsistent with established beliefs, you should start questioning the related assumptions and look for high-impact optimizations to unlock.

HEROIC Approach

What happened?

Why did it happen?

So what if it happened?

  • Have a hit list – Which Metrics, reports, segments can you investigate?
  • Evaluate data and its context – Is everything you need collected and usable? Get context by asking customers, experiencing the site yourself and learning about business changes, CI, social media, etc.
  • Recognize opportunities – Do explorative data analysis
  • Obtain deeper insights – What could lead to to your observations? List possible factors for further analysis; ask “why?” five times
  • Inspect monetary value
  • Choose the best options

Segmentation tips

  • Build best and worst segments
  • What are the majority and minority segments?
  • Which are below and above site average?
  • How do they behave before and after an event?

Web design heuristics

  • Put important CTA above the fold
  • First impressions are important – you have less than 10s
  • Keep it simple and focused
  • Have a consistent navigation
  • Communicate trustworthiness

If you fund an insight, asks what its monetary value is.

Change requires action. Action is dependent upon your company recognizing that a significant gap exists between the current state and the ideal state.

Top six reasons for manager resistance to change

  • Loss of power and control
  • Overload of current tasks, pressure, resources
  • Lack of skills and experience
  • Fear of job loss
  • Disagreement with the new way
  • Skepticism about the need for change

Resistance points in WA

  • I don’t get it
  • I don’t like it
  • I don’t trust it
  • I don’t trust you

1. Know the Audience

First of all, you want to make sure your insights and ideas are being shared with the right audience, the people who care about the topic you’ve analyzed and who have the authority or influence to pull the trigger on your recommendations.

Even if you don’t know everyone who will be joining your meeting, you should be able to obtain intelligence on their positions, backgrounds, pain points, and so on from other attendees or contacts.

As part of that discovery process, you need to determine which members of your audience are the decision makers and influencers for your recommendations. Within the group of influencers, try to determine beforehand who your allies and adversaries will be;



  • Responsible for achieving various business goals
  • Want to know, how they can achieve their goals
  • Want to know the consequences of your recommendation
  • Support: solutions to key business problems, new opportunities, better customer insight
  • Resist: Reduction in power, risky endeavors


  • Focused on hitting targets to increase their budget
  • Data-driven marketers can provide very valuable insights
  • Start using small wins to help to show the pros of the data-driven approach
  • Support: Marketing optimization, insights, improving marketing efficiency
  • Resist: Reduction of budgets, shift of resources, increased workload

IT Professionals

  • Lots of workload, have to prioritize
  • Often rigid processes and cautious
  • Support: low effort projects, have enough time, don’t require IT support
  • Resist: Unfamiliar technology, redundant technologies, problems with security/privacy

UX Designers

  • Often open but need to understand the advantages of data
  • Support: better UX, insights, streamline design process
  • Resist: Deviance of usability best practice, restrictions on creativity & flexibility
  • Product/Content Owner
  • Cares about how their product/content is performing
  • Support: insights, better performance
  • Resist: Creation/modification of content, decrease perceived content effectiveness

Decisions are influenced by

  • Assumptions: Try to make clear statements, so that people to fall back on their assumptions
  • Information: Get as much as information as feasible, esp. customer surveys, etc.
  • Emotion: There are two kinds of emotions, one kind are relevant, the other are irrelevant, e.g. car broke down => doesn’t like recommended action
  • Motivation: Self-interest is often the most powerful factor
  • Cognitive biases: e.g. bandwagon effect (we’re doing it because all are doing it), confirmation bias, recency effect, sunk cost fallacy, wishful thinking, etc.
  • Social Dynamics: Peer pression, groupthink and group polarization

2. Nail the Message

Your audience hasn’t spent hours upon hours in the analytics data like you have. They may not be aware of all the upward and downward trends affecting the online KPIs, they might not fully grasp the meaning of all the different web analytics reports or terms, they might not remember all of the insights that you’ve gleaned from past analyses, and their knowledge of statistics might be a little rusty.

Make your ideas stick:

  • Simple – one to four key ideas
  • Unexpected – Surprising facts
  • Concrete – Make your ideas clear
  • Credible – show details to make your idea seem more real
  • Emotional – why should they care about your insights?
  • Stories – anecdotes work often very well; compare “what is” with “what could be”
  • Always have a backup which includes your analysis
  • Data visualization matters
  • Use personas and tell the story of one persona
  • You can use a villain aka. benchmark

Great analytical insights that are not communicated effectively are as good as wasted. Just as much effort needs to be put into communicating your insights and ideas as formulating them. By monetizing your recommendations, simplifying your message to its core, leveraging best practices in data visualization, and crafting a compelling story around your data, you’re giving your message a fighting chance to resonate with your audience.

3. Stay the Course

In many cases, the discussion is the most important part of your whole presentation. This is where your decision makers and key influencers determine the fate of your analysis work and discuss the course of action.

You’ll learn about strengths and weaknesses

You’ll better learn about positions of the stakeholders

Support can get stronger

Retention will be stronger if your audience is not just listening

You’ll get more domain knowledge

Before the decision is finalized, you’ll want to stay involved in the following ways:

Be amiable to answer questions

Performance quick follow-up analysis

Mediate disagreements between stakeholders

Follow up key stakeholders

Testing isn’t necessary for every optimization, but it can provide that extra push needed to propel hesitant decision makers forward when the road isn’t entirely clear.

The real goal of testing is to build upon the insights you’ve gained from your analysis and to approach testing in an iterative, efficient, and disciplined manner.

Know what your primary success metric will be for each test and use it con- sistently through all of your tests.

In the initial stages, you’ll want to focus on the safer, smaller quick wins in order to build your confidence and momentum. Once you’ve gotten some successes under your belt, begin focusing on bigger opportunities that have more upside but also more risk.

Whenever you receive a thumbs down to your recommendations, seek to under- stand the reasoning behind the final decision.

Although you may not be able to completely control the execution of your recommendations, you’re still in a strong position to contribute to their success.

Verify that the teams understand the objectives, requirements and estimated value

Follow up on the progress

Be available to answer questions

Ensure proper tagging

4. Close the Loop

Do a final analysis that isolates the actual impact of your recommendations.

If the difference is significant between your original estimate and how the optimization is trending toward that target, you’re going to want to understand why.

Regardless of the results, the team that approved the optimization will appreciate your diligence in reporting back on its progress.

By showing executives the impact of their decisions, you’ll be able to build further credibility with them, credibility that can pay dividends on future projects.

In order to win greater recognition within your company, you’ll want to evangelize the successes of each team that works with you. By promoting their success, you’re not only cementing your relationship with the team, but you’re indirectly promoting yourself.

Analysis in Action

  • Acquisition: Marketing campaigns, referrals, organic, difrect, etc.
  • Site Interactions: What are they doing? Which content? What Microconversions?
  • Conversion Process: Funnel & order process
  • Visitor Value: Which customers are most valuable?
  • As a web analyst, your job in the Acquisition zone is to analyze which approaches are working and which aren’t.

Create a daily or weekly “movers and shakers” report to capture significant changes in traffic levels to different pages or site sections.

■ Objective-based. How can my company get visitors to do more of what the company wants them to do?

■ Customer-centric. How can my company help visitors to achieve what they want to do?

And yet, being data-driven as an organization is only half the equation. Very few organizations pay much attention to the other half: being action-agile

Then when I thought it could get no worse, the last manager we interviewed shared how a broken payment option in one of the company’s online stores was actually causing all of the items in the visitor’s shopping cart to be removed. Even with the severity of this problem, the web development team took nine months to fix it.

The full benefit of being data-driven won’t be realized if an organization can’t act or respond quickly to the insights found within the data

Organizations need to ensure they’re willing, capable, and ready to optimize their online initiatives.

Empirical evidence shows that being data-driven isn’t merely a good idea but actually distinguishes high-performing firms from their competitors. For example, research in 2011 by MIT professor Erik Brynjolfsson and others revealed that data- driven decision making led to a 5–6% increase in productivity and profitability.2

In fact, becoming more data-driven is a cultural change for many organizations that will take to time to nurture. If your organization’s leaders aren’t onboard, you’ve identified a key obstacle that needs to be addressed.


  • Clarify your company’s online strategy and priorities first,
  • Be strategic in which areas and KPIs you focus on.
  • Use your cross-functional knowledge to bridge teams and gaps.
  • Start building momentum with small wins that can lead to bigger successes.
  • Lead the data-driven change at your organization, and inspire your organiza- tion with your insights, ideas, and actions.

Write a Comment


This site uses Akismet to reduce spam. Learn how your comment data is processed.