I Wish I'd Known That. [Digital Analytics Edition.]

Unravel Let's start off the new year with lessons learned from a tough life on the front lines of trying to make the world a smidgen more data-driven.

This post is a collection of six things I wish I knew before I started my career in decision support systems (of which web analytics is just the latest incarnation). These lessons might have made some goals easier to accomplish, some frustrations easier to avoid and some salary jumps easier to come by.

Perhaps you are just starting out, perhaps you are in the middle of your professional journey, or maybe at the end. Regardless of where you are I hope these six lessons help speed up your journey, avoid the mistakes I made and achieve success sooner:

#1: An obsession with tools & implementations will kill you.

Everyone wants the perfect tool that bounces rates in real time while computing multi-channel attributable impact of an email sent to grandma via Facebook based on competitive intelligence gleaned from TV watching behavior of customers with lifetime value greater than $358 and FICO scores of 700 or higher.

Get over it.

The smallest part of my success, and yours, will come from having the prefect implementation of the Omniture Marketing Suite or Google Analytics.

10% of your time should be spent in implementing tools, not 15 months with an eye towards analysis in the middle of 2012.

You can win with Omniture or WebTrends or IBM or Google. Stop switching tools!

A majority of your success will come from following the 10/90 rule, hiring smart analysts and then ensuring their work day is optimally structured, weaning your management away from faith based decisions, and ensuring you have a clear line of sight and that you are following the process outlined in the web analytics measurement model.

It took me a while to realize that 100% of what makes a successful web analytics program in any company has nothing to do with tools for the first five years. After five years of success worry about tools.

Next time someone from Google Analytics or Omniture or WebTrends calls you promising to make you a data driven organization. . . tell them to suck it.

spectrum_of_success_for_a_career_in_analytics

#2: Your magnificent analytics skills are tertiary to your success.

In our industry we believe that you have to be good at creating pivot tables in Excel. And using Site Catalyst. And possess at least some knowledge of the fundamentals of statistics. And create box plots and frequency polygons etc., etc.

It will surprise you that pretty much all of your success will not be sourced from your ability to deploy the above skills. Rather it will come from two surprising abilities: 1. Your business savvy and 2. Your soft skills, your EQ (emotional intelligence quotient).

Your real output is not the data; your real output is a set of decisions the Senior Leaders can confidently make to drive business priorities. Recommending actions requires an ability to understand business strategy, marketplace dynamics, some complex situational analysis, root cause identification, and, perhaps most importantly, an ability to identify the right business question. You don't necessarily need an MBA, but if you can't have a intelligent business analysis conversation you won't be effective.

Your real impact comes not from providing pretty pie charts from a complex Discover2 query. It comes from your ability to deal with multiple psychological personalities, from being a warm and friendly person, from constructing relationships across the aisle, from your ability to speak confidently and persuasively about hard decisions and brilliant insights, and disarm people with your charm and un-arrogant demeanor.

If you want to be a great Analyst who has an ability to have business impact with data, work on the above two skills.

[I am not saying analytical skills are not important. They are, see #6 below. But without the above two you are not going anywhere, no matter how smart you are. Prioritize.]

#3: Obsess about Outcomes.

Every other channel is so data poor that when I got into web analytics, like everyone else, I was ecstatic to see very specific customer behavior like the number of Visits! Man that was awesome!

Then I saw page views. And then time on site and unique visitors and browser types and % new visits and I became one happy data-shoveling monkey.

Here's the amazing thing: everyone else who got all those reports thought I was quite simply magnificent. After all I had all this, as they put it: "OMG data!"

That only lasts a couple of months, six at the most.

Then no one cares because no one is any wiser and the business still stinks at the web.

I remember when I first created a simple spreadsheet with 12 rows that just summed up the revenue made by the divisions of the company (even that simple task no one had bothered to do) on the web. The CMO was shocked at how big the number was. The next day he approved a job req for a VP for the Web position.

Great lesson.

Companies care about money, non-profits care about impact, governments care about costs reduced. They all care just about outcomes.

Here is the only way to succeed at web analytics: Identify the business macro and micro conversions. Torture everyone to identify economic value of non-revenue micro conversions. Obsess about this analysis: What's the economic value generated by the company? Why?

Your peers will think you are the embodiment of the messiah. Your boss will listen to your every recommendation. Your CEO will invite you to a private dinner with her/his family.

[Here is my personal recommendation in this context. Obsess about the outcomes for your customers. Use Kissmetrics or 4Q and measure primary purpose by task completion. Buy some sessions with UserTesting or Loop11. Feel the pain of your customers. Identify with their failures and hug / stalk as many people in your company as you have to in order to fix things. Deliver the outcomes your customers want and you'll ascend directly to heaven when its your turn.]

#4: Be pragmatic.

I learn this lesson every single day.

At the end of a recent keynote recently the first question to me was:

"Sure you can optimize the page and campaign based on clicks and on amount donated. But how can you optimize it for the fact that the person voted in the election?"

This was in context of a President Obama A/B test.

I can't measure if the person voted.

I can measure that they came. I can measure that they signed up for a lead. I can measure that they donated and which email campaigns caused them to donate actual hard earned dollars. But no I cannot measure the offline go into the vote booth action.

I ended my answer with:

"If after signing up and donating money to Obama the person voted for Sarah Palin then you know what? I am perfectly okay with that. I did great with my online campaign!"

Be pragmatic.

You will always always, always, run into people who want to achieve the impossible on day one and push you to do the same. Who refuse to fix the high bounce rate on top landing pages. Who want to measure Engagement as defined by x time y divided by z to the power of m with the resulting sum multiplied by n and subtracted from o. Why? Maybe because they are over-achievers. Maybe because they are bored with the "mundane." Maybe because they have no idea what the heck they are talking about. Maybe. . . well who knows.

It is your job to be pragmatic.  Don't focus on what should be measured in a perfect world of persistent nirvana. Focus on what can be measured in a world that is imperfect and needs improvement now. You already have data, a lot more than any other marketing channel on the planet. Use it. Make love to what you have. Produce beautiful analyses tied to business priorities. Get people to take small actions every day and some big ones every month. Go back make more love.

Achievable victories, every day. Aggregation of marginal gains!

Don't get suckered into the impossible. Not because being ambitious is bad, but because putting points on the board matters a lot.

 

#5: Embrace agility, nimbleness & a portfolio strategy.

If I am not mistaken, for the first two years I did nothing except live in the world of clickstream analysis (with WebTrends and ClickTracks). I did some of the things recommended above (others were yet to be learned). My continued gainful employment indicated that there was some appreciation for the work. But I was still dissatisfied because there was not more slashing and burning of sites going on, not enough Executives had moved beyond using faith.

It took me some time to figure out that clickstream data left far too many questions on the table. Even though we had lots of data, we were information poor.

That led me to my first online survey. Failed. Tried again. This time it worked better. Suddenly I did not have to guess why people were dropping off like flies at stage two in the funnel. They were telling me! I knew the conversion rate for search was 2.74% but now I knew it was so low because people were heading off to Amazon and Costco to buy our product (and that was okay). I knew so much more.

I remember the first time I logged into HitWise. Total data-gasm. I could not believe there was so much to learn from our competitors that so easily helped us identify gaps in our marketing strategy. (We had not even heard of four out of top five companies getting traffic for our biggest head keyword!) It helped us find new geo's to target, new affiliates to target and more. Think of how stupid we were to just focus on website clickstream data.

Then came Offermatica. Man that was awesome. A/B testing! Finally moving beyond HiPPO's trampling the user experience. (Rather than Deciders, they became just one, of many, voices on the table powering an ideas democracy.) Also finally what a fabulous way to try ideas that came up from reading all that voice of customer.

Such a valuable lesson learned (one that would end up as two books!).

Web analytics is fundamentally about using multiple tools, because the questions we have to answer are far more complex than we are used to. It is critical to develop an internal ability (and fortitude) to be agile and nimble. Use the right tool to answer the question it is good at, and then move to the next tool and then the next one because it is better at this other thing. Web Analytics 2.0.

If all you know is WebTrends, or if all you do is spend your day with Omniture, or if your face is tanned from staring at Google Analytics, all day long you'll have a short, fruitless career in this field.

#6: Malcolm Gladwell is right, it takes 10,000 hours.

I don't think I ever appreciated what it takes to just stay current and, in hindsight, never comprehended what it takes to become good. I mean really good.

Not to be overly dramatic but. . . blood, sweat and tears.

Most blogs die in 30 days, most Twitter accounts are full of crap and have few followers, most of us never read books, most of us rarely curate a really good list of RSS feeds and then read every post, most of us will never engage in a meaningful online debate, most of us will not start a website and care and feed it and implement 15 tools every single year purely to learn and push the universe known to us, most of us never consider taking a class or two to learn new skills, most of us refuse to work a few hours extra every week, most of us refuse to experiment with what makes us uncomfortable.

And yet it takes all that to be good at what you do.

Do the job you were hired to do as well as you possibly can, regardless of whether it is a dream job or not. Then in small and big ways, figure out how to spend an extra five hours a week on you. Just five.

Your company is not going to give you this time. Your spouse / boy friend / mom might not even give you that time. Your God might be against it.

But you'll have to figure out how to watch a little less TV, spend a little less time on social media, a little less time on dinners, a little less time watching movies on planes, a little less time launching attacks on attacks, a little less time on parties, a little less time in meetings, a little less time in the bathroom, a little less time. . . something or the other. Half an hour stolen here, half an hour stolen there, invested in learning and doing and failing at the thing you want to get good at. 

Else accept that you'll be an okay Analyst, yet another non-relevant blogger, a Twitter blowhard, a winner of promotions via political machinations rather than adding actual value, always a conference attendee not a case study, someone on the professional train to somewhere convenient to others and not you.

Not that there's anything wrong with that.

Work is not everything.

But in the context of work, know that if you want to get good the path goes through putting in the 10,000 hours of hard work. Even if you just want to get good at what you do you'll still need to put in hours and hours of effort. What's more you'll have to be self-motivated because no one can want it for you, you'll have to want it for yourself.

And yes, yes, yes, there is more to life than professional accomplishments, but that's not what this post is about! [See: Nine Rules To Work / Live By]

For me trying to get really good is a journey not a destination, one on which I am still in the early stages.

There you are, six things I wish I knew before I started my career.

I hope you'll add a pinch of passion and a dash of daring and that your journey will be rich, rewarding and resplendent with glorious achievements.

All the best.

Ok it's your turn now.

In context of digital data analytics what are some of the lessons you have learned in your professional journey? Do these lessons resonate with you? What might I have missed that you would like to add?

It would be delightful to have your comments, feedback, perspectives & critique.

Thanks.

Comments

  1. 1

    Thanks for sharing.

  2. 2

    Hi Avinash

    Any plans to visit Australia?

    We've definitely got quite a small Analytics community, but I really think you could help get rid of some of the outdated ways of thinking that persist over here.

  3. 3

    And in the end the success you take is equal to the success you make

  4. 4

    Brilliant post, sir. My favorite line? "Your real output is not the data; your real output is a set of decisions the Senior Leaders can confidently make to drive business priorities"

    All too often, as you've pointed out, we create gigantic bundles of data that we can't wait to "thud" on the conference room table at our next meeting.

    We think it will really impress our colleagues and get them just as excited as we may be. But really, it just puts them to sleep. And they all just sort of want to know "ok, what's the bottom line for us here?" Despite the access to data, we must exercise discipline and sift it down to only the nuggets that truly are worth chewing on.

    Glad I found this blog! Great stuff. I look forward to subscribing.

  5. 5

    This is such a wonderful post.

    In January, I started my journey of learning web analytics. I put together a list of books, youtube videos, blog posts to read and internalize. In fact, I took the day off of work to work on this list of resources. I'm actually sitting in a coffee shop with Web Analytics 2.0.

    This post is a great read for someone like me.

    Do you have a list of books for recommended reading or must watch videos (I'm watching every single one on the googleanalyrics youtube channel)?

    Happy Monday!

  6. 6

    Avinash,

    As a new hire in the digital marketing field, Lists like these are an invaluable resource to me. I can see the usefulness in some of these tips already a week into my professional career. So many people are obsessed with the different tools that can be used and their cost/benefits, I'm already finding myself talking to co-workers about the great things google analytics alone can do.

    Regarding #6 – while I think this is completely true and is something to work towards, I would love if you could share with me some of your secrets for staying vigilant in the ongoing process of becoming truly great at what you do. I think I'm lucky enough to have the passion for what I'm doing, so the continual learning is fun That being said, its always easy to leave work at work and pursue other personal endeavors that may not be as beneficial in the long run (hello sitting on the couch watching the same sportscenter 3 times in a row). What are some of the ways you devote yourself to pursuing excellence as much as possible? This was a great post, thanks for your insights

    Nick

  7. 7
    Colleen Collins says

    Hi Avinash,

    As an inhouse SEO, my boss constantly encouraged me to give him reports that drove solutions – with actionable steps to drive change. For the longest time, I was so engulfed with the data, the niche interesting points, and the overarching ideas that would improve the website. In hindsight, I see how the best ideas (driven by data) mean nothing if there is not a clear, direct, simple plan to implement them. Your report that lead to a VP hire proves this. It isn't dumbing it down; it isn't dishonoring the data to provide a clear and simple report honoring my analysis and web metrics!

    It is a hard lesson to finally learn, but as you wrote – a necessary understanding to succeed.

    You are right too – that we have to push above and beyond – reading, writing, contributing, and learning. I think about these things a lot – but now I am making steps to do them (starting with a daily RSS feed). :-)

    Thanks for your guidance, inspiration, encouragement, and honesty.

    Colleen

  8. 8

    i wish i could get an entire week of vacation to read all your posts.

  9. 9

    As always, a great post.

    One thing I might expand on is not only identifying each of the macro and micro conversions throughout your business, but having a general understanding of the jobs that influence your site's performance. Understanding the decision process of a paid search manager or merchandiser helps guide you towards delivering the most relevant, actionable metrics & insights.

    Sometimes we become too focused on driving our career in analytics we forsake continuing to stay abreast of whats happening in the rest of the digital world (or even offline O.o).

    Thanks again for the post, I have notes a plenty!

  10. 10

    Pure genius: "Torture everyone to identify economic value of non-revenue micro conversions."

    But your advice is not just for web analytics. It applies to many things. Take the garden-variety sales process/sales cycle of 22+ steps from awareness to repeat buy. How many sales or marketing people know the economic value each time a prospect moves from one step to the next?

    What would happen if they did?

  11. 11

    You mean there's no easy path? Get outta here! I thought once you installed Google Analytics you had arrived in the land of milk and honey.

    Seriously though, thanks for yet another post for those in the trenches. I don't know how many other jobs require you to overcome deeply entrenched ways of thinking just to get to step one in a long road (maybe that's why it was so easy to move over to online marketing from project management?).

    It is a lot of fun though. Thanks for the words of encouragement :)

  12. 12

    Avinash, great blog post. I could not agree with you more on #6.

    One of the biggest lessons I have learned is that strategzing before implementation is key. Before, I use to just attack a problem or issue head on without any strategy and that caused wasted time management and a large room for error.

  13. 13

    Wow! I really needed this post this morning. Especially section #6.

    Often times I find myself just turning off after work when that is the time to actually experiment and go my own way.

    Thanks for the motivational kick!

  14. 14

    Wise words.

    Thank you very much for sharing. I definitely need to work on my EQ!

  15. 15

    Before implementing our ideas and plans we need to recheck them twice. It can make our of time wastage.

    Thanks Avinash for sharing the info…

  16. 16
    Manish Tripathi says

    First of all Avinash. thanks for the wonderful post. It was not only informational but I truly found it very inspirational. Thank you so much.

    My question is in your above post, dont you think that your 1st and 5th point contradict themselves.?

    In the first point you talked about not going for many tools and stop obsessing for more than one tool, while in the 5th point you have stressed on knowing mulitple tools is very helpful.

    Both of them look contradictory.

  17. 17
    Alice Cooper's Stalker says

    Avinash,

    Great list!! Lots of good gems of information within each of your six.

    Here are a couple of my 'Wish I had knowns' although they are a little more detailed….

    -Trying to maintain and support an in-house hosted web analytics solution can suck the life out of you and significantly impede your ability to deliver real value. It can be a major time and value drainer.

    -Each tool (and sometimes report) has it's own nuances. It would have been helpful to have a comprehensive list of those nuances, thresholds, limitations and quirks. I guess you learn as you go.

    -There are resources like Occam's Razor, the Yahoo Web Analytics message baord, WAA, Web Analytics Demystified and other blogs that can help you get a jump on those 10,000 hours that Malcolm Galdwell talked about. As you indicate, you need to put the time in to get the value.

  18. 18

    "Companies care about money, non-profits care about impact, governments care about costs reduced. They all care just about outcomes."

    So very true!

    Thanks for this post Avinash – it's quite helpful for someone just getting started.

  19. 19

    Great post Avinash – Agree on all points!
    Thank you!!!

  20. 20

    Great list and so many applications for it – recent grads, digital marketers, and of course, web analytics professionals.

    As several have commented above – much of this advice spans well beyond the web and analytics.

    Especially this quote "Companies care about money, non-profits care about impact, governments care about costs reduced. They all care just about outcomes."

    If you can start each day with that in mind it will help drive decisions and keep focus on what's really important.

  21. 21

    One of the biggest lessons I've learned, thanks to you, is to always ask the "Big what and/or why"? Why are we measuring this and what will this this metric tell us? It is very easy to drill into data (there is tons of it), but at the end of the day what *outcome* is it feeding? What business goal / decision is it helping us or our clients make?

    These questions fuel continued learning for me because each client has a different what/why.

    Thank you so much for always sharing your insights and lessons learned.

  22. 22

    I did 3 years "clickstreaming and measuring outcomes". That was fine (nice to see that Avinash spent 2 years with that!), but as you can imagine, we didnt improve too much between years. Then some organizational changes happened in our web team, so i started asking myself some questions and i have found this blog.

    Since October i have read about 20-30 articles from this blog, in end of November ordered & finished first book from AK and im actually in the middle of second book (my goal is to read and understand 20 pages per day, it takes me about 2 hours daily, one in work and one in home in the evening works best for me).

    And thats it, its so simple, after 2 months of extra educating yourself (it means approx. 60 days * 2 hour daily = 120 hours) i feel much more comfortable and savvy, if i speak and communicate with my boss or colleagues about online things :)

    My only problem now is, that we actually dont have more "web people" in the company, so words like "pragmatism" or "evolution works, not revolution" are one i have in my mind every single day now, and thats the biggest challenge, as no one actually cares about clicktream data and rarely cares about outcomes data, which is ofcourse sad! :)

  23. 23

    Avinash,

    Great post!

    The timing of it is perfect. I am new to Web Analytics and this is just the information I need to get my career going.

  24. 24

    Great post, Avinash.

    I think the issue with #1, is it depends on who you talk to, what person in the organization. A healthy level of obsession with the tool and implementation, in my opinion, is a good thing.

    I can't begin to tell you the number of times I've been part of an analysis, only to find out after much digging, that the data was completely useless, due to a poor underlying foundation. This can be remedied by being on top of the tools/implementation.

  25. 25

    I think the most important thing I've learned is that the way to approach business and analytics is to *allow your instincts to tell you what data you need.* I think people lose themselves in the shiny offerings of their data packages and swim around in a pool of non-actionable data, surfing analytics like they surf the web, rather than utilizes the data and creating their reports *after* determining the specific question they are trying to answer.

    As perhaps an extension of #4, I've also learned that a large number of business managers, owners, execs, etc. often have very little idea of why their customers behave the way they do, and why their business is successful or unsuccessful to the degree that it is.

    The businesses that I've helped the most are the ones that are just profitable enough to be comfortable, or who are experiencing slow but stable growth, because those are the ones that are lazy about looking into why and capitalizing on the areas in which they are succeeding, assuming that if they stay the course and maintain the status quo, they'll continue moving in the same direction. They could be finding major wins and predicting how things are going to change, rather than reacting after the change happens.

  26. 26
    Justin Knightley says

    Thanks Avinash, interesting reading – as ever. I wish I could simplify things better (for top management).

  27. 27

    Another great post!

    I still consider myself new to the web analytics world, but throughout all my years in analytics, one important thing I've learned is the question "Why". And I don't mean in the sense of the data, but when a HiPPO comes and asks you "How many visitors did we get in August?". That is a request for DATA. I don't just simply ask "Why" they want it, but engage in a conversation and ask them what they are hoping to learn from the data – what BUSINESS question are they trying to answer.

    Most times you'll find the data point they asked for is not going to answer their real question. You touched on this above, but I can't stress it enough. An analyst can save themselves a lot of time and energy by getting to the route of the question – and can also avoid becoming a report/data monkey!!

    Thanks again for all your great work and insights!!

  28. 28

    @ Keith

    That is so ridiculously true. Higher up's will often ask for something without telling you why, and if you dig deeper you realize what they think they want won't answer their question…

  29. 29

    Hi Avinash,

    Great post! I wish I have the time luxury of reading all your blog post as I am new in the Web Analytics field. Cheers!

  30. 30

    Tim: I love that! I really do.

    I should have ended my #6 with that quote. Will certainly do so in the future!

    Andrew: You are certainly not the tool!!

    There has to be a balance. You can't be up on everything all the time. Setting aside some time, say 10% of your time or 5% (you choose) to experiment with "new stuff" is always of great value. Pushes you beyond your current normal, forces you to learn.

    PS: To contradict myself… someone on Twitter remarked something like: "Don't forget that Omniture is not the tool, your brain is. Use the right tool!" I like that thought. :)

    W_CARR: I think between Web Analytics 2.0 and reading a few different blogs (see list on left nav on home page) I think you might have things well covered. If you are just starting out getting some training / book reading about your specific clickstream tool can be helpful. For GA for example the latest books by Brian Clifton and Justin Cutroni. Snag the Cartoon Guide to Statistics (http://goo.gl/mxuhT) if you don't have a background in that area. Other than that (when you have time!) marketing and psychology books are great. Don't Make Me et al.

    Nick: It is hard to share "here is how I stay up to speed" in a pithy comment. But here it goes:

      1. I am constantly pushing my work at work to be on the bleeding edge. So work with really smart people, push the team to constantly explore new things, occasionally be very critical (approach bitch-fest zone) for what we currently do.

      2. An expansive, diverse curated RSS feed. There is gold in blogs if you don't follow the distracting ones.

      3. Practice a lot. I am lucky to have this blog where I dump and try and fail at a lot of tools.

      4. Loads of conversations. My speaking engagements page should give you a clue that I am out and about a lot so seek out a lot of people not in our industry to talk to.

    Hope this helps a bit.

    Dave: That is really great advice.

    Far too often we get stuck in just website analytics (incoming traffic data and site / page performance). Having some understanding of PPC and SEO and Affiliate Marketing and Email Marketing and Social Media Marketing and… some of the top marketing techniques is absolutely critical for any good Analyst.

    First because you'll get an awesome understanding of 1. what data you are missing 2. how to track these key marketing techniques and their complexity better (and no it does not happen magically when you slap a javascript tag!).

    Second because you'll just becoming a better Marketer and student of the web's evolution as an acquisition and conversation channel. Then your numbers will make a lot more sense for you for what they show and what they don't.

    Thanks again, really great advice.

    -Avinash.

  31. 31

    Manish: It might seem like like conflicting lessons. Allow me to clarify.

      #1 is very much deeply centered around a near deadly obsession large web analytics consulting companies, and large actual companies, have (and I had) around clickstream tool implementation. 90% of their time (and billable hours for consultants and waking hours for Practitioners) are spent there and then subsequent data puking. My call to arms if you will is to focus on analysis and the web analytics measurement model and… informing business users of what to do. Maybe even 90% of the time. :)

      #5 is centered around expanding your (and my) known universe beyond clickstream because that is just 20% of what you need to help your company make the decisions that make the company money and website visitors happy.

    To summarize… pull oneself out of the sink hole of clickstream analysis and never ending implementation of ever more complexity in Omniture / WebTrends / IBM / GA and do actual analysis and power that analysis using multiple types of data sources (most of which you don't have to implement, you just use 'em!).

    Hope this helps. Thanks for the great comment.

    pdxKate: I love it! Two of the most important lessons that we could learn in this space.

    There is one more nuance to this I have learned is from my good friend Mitch Joel: "Always why before what."

    That is an awesome guidance to all of us, Analysts. Why are we doing what we are doing as a business? What should we measure to understand the success of what we are doing as a business?

    Eelke: I am genuinely impressed with your commitment. I am positive that with that level of investment that you are making (all on your own time!) that you'll be able to, soon, find ways of making your executives care more about data. You already know the mantra: It's all about outcomes! :)

    Jason: I hope #1 does not come across as a license to do shoddy implementations and move on to analyzing data. My point was you got something already implemented well enough, now get out of the constant sprop and evar and vista and saint rules and events and custom variables and code scans and data warehouse etl etc and do some work. Do actual analysis!

    Perhaps this decision making matrix should have been there for clarity:

      + If 90% of the Consultant / In-house Analytics God's time is spend on capturing, cleaning, reimplementing then you will die an ugly death but, with clean data. To avoid spend 60% of your time / money on analysis.

      + If 90% of the Consultant / In-house Analytics God's time is spent on analyzing data to make recommendations that drive the business, but these decisions are 25% imperfect because of wrong code or missing sprops and variables then you'll be publicly embarrassed but will still be loved for all you have accomplished and allowed to live another day. Take 20% of your time and fix imperfections.

      :)

    A clickstream tool implemented well enough is important. And an obsession to ensure it is well enough is justified. I completely agree with you.

    Keith: That is an incredibly valuable lesson to learn. (So valuable that I had a dedicated blog post on that: Rebel! Refuse Report Requests. Only Answer Business Questions, FTW)

    It is much harder to ask the Why question and then deal with the consequences. As you point out it is much easier to data puke. But for all the trouble that "Why" is, the rewards are 100x. So we must persist

    -Avinash.

  32. 32

    Hey Avinash,
    Cool Post, I like the part of EQ, emotional Intelligence. I remember when I read that book from Daniel Goldman.

    Particularly in my position I have to balance a creative side with an analytic brain, and one of the big challenges is not the technical side; but how to segment the tidal wave of information GA gives me and to present the correct data, in the right way, at the correct time to the right people. Its fun/frustrating but such is the way of an Analytic ninja in training.

    Thank you for helping us Avinash, Rock on \m/_

  33. 33

    Thanks AVinash i got really touched by your quote: "For me trying to get really good is a journey not a destination, one on which I am still in the early stages"

    It makes me keep doing what im doing and put an extra effort to acomplish my objectives.

    Thanks again, really fan of you …from PERU

  34. 34

    Excellent post as always Avinash – thought provoking and inspiring.

    One of my faves is "Obsess about Outcomes." After becoming a reader of your blog I always ask myself "So what?" when creating/viewing a report…am I just looking at data or seeing insights?

  35. 35

    Hey Avinash! My most important lesson in web analytics (that you haven't already highlighted) is: boldly go where no man has gone before. Don't be afraid to be different, to be wrong, to learn – as long as you love doing it, and believe in it, it doesn't matter.

    Another thing:

    "Don't let them promote you. Don't let them transfer you. Don't let them do *anything* that takes you off the bridge of that ship, because while you're there… you can make a difference." – Captain Kirk (Shatner)

    Yes, Star Trek applies to everything.

  36. 36

    Hey Avinash! Great Post!

    One of my lessons learned is to define business goals properly. I still see businesses try to define complex KPIs while they don't even understand the basic ones. It is critical to understand what being measure and why it needs to be measured. Even if you can spend 10,000 hours on looking and analysis the data but if there is no real need, it is a waste of time.

    On the other hand there should be a balance of data and common sense. There are companies who too focus on data and not react to the market fast enough. Data is the result due to an action or decision being made. So the company should define when to use data and when shouldn't when it come to real life business.

  37. 37

    Hi Avinash,

    Thanks for a nice post. It was a nice start in digital marketing strategy. After a long years experience in B2B, Just i have entered into B2C. It's a awesome tips to learn more things here in your post.

    Thanks for sharing! i loved it all the points in my mind. :). i will post it (URL) in my blog/twitter.

    Thanks
    Thanaraj

  38. 38

    Hi Avinash,

    Thank you so much for this great article :)

    Analytics help me to listen more before judge !

    Thanks
    Abdelrahman – EGYPT

  39. 39

    Hi Avinash, sometimes I just wonder do you get insight from data or find data from your insight.

  40. 40

    Hey Avinash,

    This is absolutely a great start of my New Year 2011. I wish you Happy New Year 2011.

    I totally amaze because I don't know how you write these posts. You know people do Yoga for being healthy. And the same way, I follow your posts for being better and and better in my web analytics job role.

    I am following your blog since 2007. This has been helped me in my web analytics career like the way of life through Bhagwad-gita when I started my Web Analytics career in 2007.

    I want to thank you from my bottom of heart because your posts are really helped me to make better decisions.

    For the current post I liked mostly #6. That really helps.

    Also some one above written that 'i wish i could get an entire week of vacation to read all your posts.' but I have done that by giving the time for it.

    Looking forward for more posts and motivating thoughts.

    Thanks,
    Bhagawat.

  41. 41

    Great post. The only thing I am not if it is implicated here is the need to continually renew yourself or re-invent yourself. I'm on my 5/6 different career.

  42. 42

    Great post! I agree completely: the thing to obsess about is outcomes, not tools.

    The longest journey begins with a single step: don't wait until you have it all figured out. Just start, and learn and change as you go.

    And yes, I agree, it's hard work ..and very rewarding! Thanks for the reminder about the extra 5 hours!

  43. 43

    Hey Avinash, it wouldn't have killed you to have posted this in 1999! :)

    Just kidding. As the 40+ comments before this one can tell you, this is a great thread regardless of where you are in your career.

    My thoughts:

    I've been thinking about this blog post all day yesterday and driving into work today, and I'd like to share the following thoughts because I think that there may be others whom have read this blog and kinda thought the same thing.

    At the end of the day, if your clients / company are happy and if your company / clients are improving and benefiting from your efforts, then that's all one can really ask for. I've realized that you don't have to be a rock star like an Avinash Kaushik or like a Jim Sterne or like a Bryan Eisenberg to be successful and good at what you do – just concentrate on doing a good job using the many excellent blog posts on this blog as your tools and your education, and if at the end of the day your company / your clients are benefiting and improving, then you should be proud and feeling a sense of accomplishment and all of that, and have your "datagasms" if you wish :)

    The one lesson that I wish I'd known is to not try to be someone you're not. That comes through, even in Web Analytics. Be who you are and develop your own uniqueness.

  44. 44
    Mahesh Srinivasan says

    Great to start the 2011 year as an aspiring web analyst in reading this post. Have an obsession to follow as mush as said in this articles. Thanks Avinash.

  45. 45

    Hi Avinash,

    Thanks for this post. As someone who's younger in the field (I guess 26 would be considered young??) these kinds of posts are invaluable. In fact, I wish there was a national day for all the blogging guru's to spill what they wish they had known when they were younger. Not only are the suggestions useful, but it reinforces the flaw-filled journey it took to get them there.

    I keep coming back to #6. I've heard Gladwell's theory before, but hearing you articulate what those hours means makes it a little more relatable. I understand Joe's point above that you don't have to be an Avinash or Eisenberg to be successful (and thus don't need the 10,000 hours) – but I can't help but think, "if I don't read more blogs, read more books and spend more time getting better at what I do I'll never rise above the crowd. I'll be very average and average is so boring."

    I guess this is something I'm still working to balance.

  46. 46

    Thanks as always for the insights borne of hard-won experience and the inspiration borne of your passion for excellence. :-)

  47. 47

    Brilliant advice.

    Sometimes people forget the 90/10 rule and that causes more harm than they can imagine.

  48. 48
    Steven Sun says

    Great post, Thank you!

  49. 49

    Since you asked for ours, here's mine (it probably more echos one of yours):

    Pick your language carefully, because you only get one first impression [each meeting].

    We tend to use a lot of jargon and focus on either technical or trade-specific issues when we relay our analysis to others. I think we do this because it mimics our own process: we "replay the tape" for people when we talk about what we discovered by outlining where we started, what we found, where it led us…eventually ending at the revelation. So we lead with our language and our process, and END with the business's language and process.

    I wish I'd always known that you should just completely flip that process around the other way. START with the business language. START with the Outcomes you're talking about, and if the idea is good enough, the "business" (maybe it's an executive, the CEO, leader of another department, etc.) will ask you for more detail. And trust that they will. If they don't, I've found I didn't sell the idea well enough. People buy good ideas, not my process.

    The worst and most uncomfortable meetings I've been in have been the ones where I explain something for 3-5 minutes, stop talking, and nobody has a clue what the hell I'm talking about. From that point on, no matter how well I do in re-explaining, it's not going to work as well as if I'd done it well the first time, because now I have to overcome the confusion or boredom I just created! So you only have one chance to make a first impression with people (to show them you can make their life better), but you also only have one chance to make that same impression and pitch EACH meeting you have with people. If the first impression is me going ballistic geek on them, it's an uphill battle from there!

    So, I wish I knew how important this was. It would have saved me a lot of ruminating about what I did wrong and how I could shed my nerd reputation and show them I meant business from the beginning.

    Killer post.

  50. 50

    Great!

    Don't you think it was worth not to knowing something in the past to create such a nice and motivating post today??? :)

  51. 51

    Avinash: to each of your valuable points I'd add this one – create deeply meaningful visualizations of what you discover from your analyses. Let your audience see + understand what you're talking about, along the lines of what Edward Tufte advocated in his seminal book Visualizing Quantitative Information.

    I recall a costing analysis I did in the mid 1980s. It took me 3 minutes to explain to our Executive Team what the graphic I'd prepared revealed about where a missing $5m had gone. No one said a word for 5 minutes. The facts were clear. What to do about them wasn't. The conversation was informed + focused in a way few others had been up until then.

    Drawn effectively, a picture can be worth far more than a thousand words.

    Trust this adds some value. – John

  52. 52

    Awesome post Avinash.

    Point 6 just hit the nail on the head for me.

    I think assuming that everyone knows what you are talking about instead of taking the time to explain your point is something that should be taken into consideration.

    Having a teachable spirit in this business will be helpful and will pass a message across better so that you don't repeat yourself every time

    My 1 cent.

    Awesome post!

  53. 53

    Thank you for your insights, I wish I was just starting out like many of your followers to have a resource like yourself to learn from. As an aging primate, an engineer and numbers guy I have a tendency to focus on the analysis rather than the drivers and outcomes.

    The main lesson I learned (#2) is that you can't rely on anyone else to advocate the value of the numbers to an organization, balance is the key and if you want to be heard you have to know the language and triggers (key words) of the decision makers.

    Looking back, I needed to promote the value of my analytic work in the terms that the "customer" could appreciate and act upon.

  54. 54

    Thx, Avinash… I second what you wish you'd known (and some of the other posters, too) – particularly the part on EQ and its importance for effectiveness.

    To the mix I would add:

    1. Data is not information – I wish I'd known that the "best" data doesn't matter if it doesn't give the audience information to answer the problems/questions they have. Conversely, rougher information (with properly disclosed data issues) can be more helpful than squeaky-clean data.

    2. A question that's old hat to me is brand new to the person asking it – Sometimes I would get so comfortable with a standard question, I would forget to take people on the journey and feel that I was being efficient in answering their question quickly. They just didn't feel listened to. And sometimes, there were new twists that meant the standard answer didn't work for their question. I wish I'd known earlier on how important it is to take the journey with every question on the table.

    Look forward to seeing more!

  55. 55

    Erica: Excellent question! The answer of course is: Both. This is why *analysis* is so important and reporting just something we have to do to survive in this world.

    When you are doing analysis sometimes you are in a exploratory mindset, you know the goals of the company and you are exploring trends and patterns in data to find insights. At other times you'll have a hunch or a gut feeling about something or (my fav) a hypothesis and then you'll don your analysis hat again and based on that you'll analyze data to validate your hunch / gut feel / hypothesis.

    So insight from data or data from insight depending on your starting point.

    Joe: A while back I had shared this advice with someone thinking about a job change. Use three criteria:

      1. Are you adding value to your company (more revenue, happier customers)?
      2. Are you being compensated optimally?
      3. Are you happy?

    Your comment made me think of this criteria. In the end that's the goal, those three things. Things I have to say, things you have provided as valuable advice in your book, things others say as their lessons, are all geared towards solving the above three things. I hope. :)

    Thanks for sparking that thought.

    Chris: That is a great articulation of #6, thank you.

    Each person should be who they want to be. But no one should be under any illusion as to what it takes to be non-average in a professional context. 10,000. Ok maybe at least 6,000 a couple standard deviations away from the mean.

    One can choose to be sucky, average, good or great. Choose and then execute from 0 to 10,000 and then desires will match outcomes. :)

    Evan: Great advice on the power of being able to communicate. Not just "communicate to inform" (data puke!) but "communicate to drive action" (only deliver recommendations with data in the appendix).

    This is of course our aspirational goal. We still have to do reporting and endure daily implementation torture by vendors and IT and do 27 tab Excel "dashboards" with technical terms. But at least the goal should be to aspire to "communicate to drive action".

    Money quote in your comment: People buy good ideas, not the process.

    I am going to use it!

    Peter: Really great advice, I am glad you added the comment. I was only tangentially alluding to this specific issue when I wrote the #2 and stressed communication & charm and leadership.

    It is so important to become, pardon this overused term, an Analytics Evangelist in your organization and champion the value of data, the value of actions you have recommended, the value of investment of the company's time and money in you (and me and others). We think producing numbers is the end of the journey but, as you have astutely pointed out, it is the first step.

    Erin: Two very valuable lessons.

    It has been my humble experience that "dirty", but as you say properly disclosed, data now is so much better than clean perfect data 9 days / weeks / months from now. I also love your idea of "taking people on a journey".

    Thanks for adding your wisdom to the thread.

    -Avinash.

  56. 56

    This is incredibly inspirational. Thank you for this post.

    I especially appreciate #6. I think the biggest (and most important?) challenge in life is deciding where to focus those 10,000 hours.

  57. 57
    digitalpadawan says

    Love your blog – I am fairly new to the digital industry and do not work within analytics itself, however, I feel that it is important to keep up to date with everything going on in the industry.

    This blog gives me a great insight into web analytics and has helped me develop my knowledge – best blog that I have read so far and will definitely by sharing with others around my organisation!!

    Thanks,
    digitalpadawan

  58. 58

    All of the content in your post was useful, but the really powerful and impacting information was at the end.

    Very good stuff!

  59. 59

    I absolutely agree on needing to look at the numbers pragmatically and with an eye on the desired outcome.

    We get presented with quite a lot of web analytics now days as justification for a business plan or idea that may need funding. In themselves they are fairly meaningless, it is how they are used that counts and that part is often not thought out.

    Just because a site has a lot of traffic, doesn't mean they are making money. The business model and how they generate the revenue aligned to the statistics is the key.

    Finally when in doubt ask. Occasionally asking customers or visitors what they were looking for, or why they didn't buy can cut through to the chase. Well constructed (not long) surveys can add light to some of the darkest numbers.

  60. 60

    Needless to say great post! One of my friends always say "You should not repeat the obvious things.." so I will do it again…

    Avinash your list and readers' comments capture the points implicitely which I am going to mention but still thought of stating them..

    I believe one of the traits of any web analytics professional should be the correct mindset. And when I say that I mean that think like a businessman and not like an employee or a marketer. Any true businessman would be concerned about the outcome and would focus only on solutions and not the problems as highlighted by you.

    Another point is "Open Mindedness" which leads to learning, experimenting and exploring which connects with your last point and many readers' comments.

  61. 61

    Great post! I must say it feels like I am reading some novel for lesiure..does not feel heavy or Jargony at all:)

    Having said that I do question the efficacy of web analysis? I still dont believe that all that web data actually tells us much about our customers! I would love to know what are other realistic ways of measuring customer satisfaction.

  62. 62

    I wish I had said number 6; very well done!

  63. 63

    This post couldn't have come out any better time than this. I was in the verge of losing my faith in google analytics, even made plans for an Omniture user training to move on, but your post sheds a very different light on the matter. Thanks Avinash, thanks for summarizing your vast experience and presenting it on a single blog post, full of yummy inspirational ideas to get back on that career development track.

  64. 64

    You are spot on in #6.

    I learned much while sitting in waiting rooms, parking lots waiting for the kids , ice skating rinks and reading every spare second available. It all adds up in terms of learning.

  65. 65

    I just stumbled across your blog this evening. I have been in the web world for over 10 years but have only recently started focusing in on SEO/Analytics. Great article but the two points that resonated the most with me were #1 and #2.

    So true!

  66. 66

    Great information!

    Again thank you Avinash for sharing unvaluable knowledge.

  67. 67

    Avinash,

    I was baptised into Digital Marketing 6 months ago with your blog and I immediately picked up your book as well. What followed is as predicted in #3 of your post: 3 months of frenzied "activity" in getting my arms around Omniture(HBX) data on the my companys website, staring into the adwords dashboard every morning and brainwashing my digital agency to use GA to track the conversion funnel ….at the end of it all I was exhausted .. there was soooo muccch data …but the actionable insights were still hard to come by..and I stopped following your blog !!

    After reading this post …I am taking a step back to reassess my approach to use of web data and analytics …thanks for the wake up call for 2011!

    Monappa

  68. 68

    I think no1 can be summed up with Marshall Mcluhan's observation back in 1964 – "We shape our tools and afterwards our tools shape us".

    It's sad that web analytics jobs focus on what toolsets you have worked with eg from a recent job ad "3-5 years as a web metrics analyst with experience using web analytics tools, Coremetrics is preferred but we will consider candidates who have used one of the more mainstream analytics tools such as Omniture or Web Trends."

    This should not be the main crux or a dealbreaker when it comes to a hire. If you've used one, you can pick up how to use another fairly quickly and really it shouldn't matter.

    cheers

    Jon

  69. 69

    Hi Avinash,

    I stumbled upon your blog while looking for resources to teach myself Google Website Tools/Analytics. I'm a recent grad who wants to pursue a career in marketing. I have great interest in technology, so I plan to apply to online analysts positions at marketing agencies. Now, the job requirements list previous experience in Omniture, which is an expensive program to get training for.

    My plan is to train myself in Google Analytics and leverage my knowledge there to present myself as someone who understands marketing analytics, and is positioned to use Omniture, given minimal technical training.

    I'm curious to hear your thoughts, in regards to your #1 point. Do you think this is a good idea? If you have any suggestions on the best way to learn marketing tools in order to be more marketable, feel free to suggest.

  70. 70

    Kristine: It is extremely important to be aware that there are four distinct career paths in Web Analytics. Each path requires a different set of choices and investment in yourself. There is at least one path where an obsession will tools and implementation and javascript will pay off handsomely.

    Please refer to this blog post for job titles, salary prospects etc for the 4 paths:

    ~ Analytics Career Advice: Job Titles, Salaries, Technical & Business Roles

    As to "any suggestions on the best way to learn marketing tools in order to be more marketable" I'll recommend the specific tips and ideas recommend in this blog post:

    ~ Web Analytics Career Advice: Play In The Real World!

    I see that you already have your website and are active on it, you already have a leg up on others!!

    All the best.

    Avinash.

  71. 71

    Little bit off-topic question, if i can.

    What web analytics solutions are typically used in banks, tags? Is there any bank who uses lets say GA in whole website, so also in client zone, where clients make all transactions?

    Im asking because actually consider one job offer and that company (bank) uses GA except client zone, where due to security its forbidden for GA, so i would like to know little bit more about "standards" in this segment.

    Ofcourse thats a question of objectives of website, i can imagine that its not necessary to know analytics about client zone activity and measure click to zone would be enough.

    Thanks in advance.

  72. 72

    Eelke: It depends.

    There are hundreds of banks that use Google Analytics (and its peer Yahoo! Web Analytics). Many of those banks also chose not to use cloud based solutions (including Google Analytics and Omniture and WebTrends and CoreMetrics etc) for the client transactions area for reasons they deem important. Many of these banks (or indeed other companies) use hosted solutions to measure customer experience (like Urchin, hosted versions of WebTrends etc).

    There are no "standards" because each company's needs are unique and their world view is even more unique.

    Avinash.
    PS: I would not agree that it it is not important to measure, as you put it, "client zone activity". Web Analytics includes analysis of the, in this case, "open website" and roi for marketing etc, but it also includes analysis of the consumer experience "behind the login" and in RIA's and Applications to ensure that the web based consume experience is the best possible. Often there is much more ROI in analyzing and improving the "client zone activity".

  73. 73

    Avinash: yes i completely agree with you and your answer helped me a lot! I asked at job interview what they use in client zone (because find GA in rest of web was simple) and got reply in style "are you crazy put something in client zone?" :-)

    So good opportunity for me and the company, if i move with my skills there!

  74. 74

    Thanks for the link! Reading that post helped me understand the various types of roles that are available in the analytics world. I'm able to narrow down my choices to two, and I'm still trying to determine where I would fit best. I want to be in an analytical environment, but analytical not in the technical/advanced stats, but more like conceptual, strategic thinking. I have fluency in technology, but mostly on the user side, compared to working on the backend/programming side. Marketing analytics or even marketing research (consumer behavior) are areas I'm considering. I'm very cerebral but I've also had leadership positions before where I've succeeded, and right now I'm figuring out whether I should go the individual contributor path or leader path.

    My solution is to take the individual contributor path (for now), further my leadership skills internally within company or externally with associations, then when I have some solid experience in my belt I'll try the business team leader route.

  75. 75
    Bernardo Doré says

    It's funny how people are alike. Excellent post.

  76. 76

    Gladwell was right in regards to the 10,000 Hr Rule. The Beatles wouldn't have been who they are if it wasn't for playing in Hamburg strip clubs and Bill Gates and Paul Allen wouldn't have been who they are if it were not for the hours upon hours of code writing and data processing in the genesis of computer programming.

    You only hear about the glory years of famous individuals but you never hear about the struggling times that they perseveared and the long hours it took to get them to where they got to. 10,000 Hr Rule is a great standard to live by.

  77. 77

    […]
    I was reading the latest blog post by one of the undisputed authorities in web analytics, Avinash Kaushik. He wrote a post titled: I Wish I'd Known That. [Digital Analytics Edition.] In it Avinash says:
    […]

  78. 78

    Avinash,

    I can't agree with this post more. I particularly liked your reference to a portfolio strategy and I think this is what separates the reporting squirrels from the analysis ninjas. A lot of insight can be gleaned from clickstream data but it is only a piece of the digital puzzle. Gathering feedback from your user community by conducting user testing and getting a sense of the competitive landscape through a Hitwise or a Compete brings together all the pieces of the puzzle and empowers your recommendations with a more informed view of the challenges and the specific steps needed to meet those challenges.

    Another great post sir! How's the photography coming along?

    Have a great day!

  79. 79

    I'm a true believer in hard work gets you everywhere.

    One of my favourite quotes in this whole post is right at the very beginning when you were talking about your career in 'decision support systems'

    If more people looked at analytics like this, as part of the overall decision making process within business they would be much more successful in their role.

    Great work Avinash to crystalise these thoughts.

  80. 80

    Great summary. Data isn't enough and can be worse than nothing at all in some instances.

    Jeff

  81. 81
    Blanchard says

    Avinash, agree with all your points on this blog except the prescriptive sermon on squeezing out those 5 hours. I think 'your' blog space deserves more than this, more specific because its been a good blog all along.

  82. 82

    Avinash, thanks, as per usual you are an inspiration.

    A year into trying to get a real analytics & optimization program going I am sitting here at my desk thinking "maybe I am doing this wrong, because it is harder than I thought it would be".

    This was a helpful reminder of the journey we're on!

  83. 83

    Thanks Avinash for this great post. I totally agree with your idea of "outcomes". Not data.

    But here in Germany many CEO's are far away from your way of thinking and acting. Therefore it is delightful to read your posts. They often give me new insights on which i can experiment and act.

  84. 84
    San Diego PR guy says

    Dittos, of course. No matter how complex some of this seems, especially to a beginner, it is worth knowing that there are an amazing number of companies — $100 million plus — who have web sites with NO ANALYSIS. none.

    So there is still lots of room for people asking the right questions.

  85. 85

    Thanks for sharing Avinash. I especially like the part about changing the way an organization thinks of site optimization. The fact that it should be data-driven and not faith-based. I can see that happening now in my org. and it is way more precious than merely being able to crank out fancy reports (which I am still learning to do every day). In my view analyst need to stay ahead of the curve and make sure that their org. doesn't fall behind at the same time.

    Great post!

  86. 86
    Aurelius Tjin says

    Thank you so so much for sharing this post. The timing is so perfect. Thanks a lot! It's worth to read.

Add your Perspective

*