Gut and Gumption - How I became a web analytics guru

It was early 2006. I had recently discovered the magic of Search Engine Marketing (SEM). It somehow brought out the best in both my numeric skills as well as my felicity with words. It was a heady space full of learning new things on a daily basis. I wrote intellectual looking articles on Search that impressed clients and in fairly quick order I had metamorphosed into an SEM guru.

A massive Enterprise SEO opportunity

Around that time, one of the largest travel companies in the world put out an RFP for an SEO partner. It was an incredibly prestigious client assignment, and was commercially interesting. The client was tech-savvy with huge and global scale.

I had to redo my SEM act, transmogrify into an SEO guru and win the business !

In the Kingdom of the blind the one-eyed man is king.

I knew very little about SEO. It was still the domain of technical geeks and webmasters who were focussed on arcane aspects of gaming Google. But I took refuge in the fact that this was a rapidly emergent field, SEM and SEO were 2 sides of the same coin, and I would "wing it" somehow. I seemed to be the only person in the room who could speak for over 2 minutes with any conviction, which reassured me.

I combined my sales skills with the domain expertise in Search, and googled my way to a impressive SEO presentation. It made sense, presented the team's credentials and approach well. We decimated the minimal competition from a few media agencies and were invited to commence commercial discussions as the selected vendor of choice. I was now officially, an SEO guru as well !

Commercial Discussions: Significant Performance Bonus

The commercial discussions went well, and we arrived at a good base fee, with a 15% bonus every quarter for 4 quarters, should we grow the organic traffic from Search Engines globally by 15% relative to baseline. This was in a situation where organic search traffic was growing at scale 20% annually. So we had to outperform the market 3X. (60/20)

So if the base fee was X, we stood to gain 60% of X purely on performance. We signed the deal and the client gave us access to their Google Analytics console. Google had just purchased Urchin in April 2005, made it freeware and called it Google Analytics. Our client was possibly the earliest and largest implementer of GA, with Google engineers flying down from Mountain View to implement and supervise the process. The implementation was quite advanced. I learned all about Custom cookies (UTM V cookies), that permit the build of behavioural profiles, long before mainstream practitioners understood what was possible. Everything looked promising,

The party begins......with unexpected twists....

We received GA access from the client. It was the first time that I saw the volume of data. The brand had over 75 global domains, and well over a million transactions a month, several of which involved multiple steps. It was analytics heaven,

But soon, there was some sort of cognitive dissonance, something that I couldn't exactly put my finger on. The data simply looked, WRONG.

The likelihood of us getting to the bottom of this was minimal, given my meagre knowledge technical aspects of GA implementation as well as SEO. But I was convinced that there was a rat.

We met the client and raised the issue. How could we agree to performance-based compensation when we were not even sure if the baselines were correct ? It seemed the wrong thing to do, and I was close to declining the assignment if we couldn't figure out the issue,

The client, a practical and sophisticated e-commerce practitioner empathised with my apprehensions. She gave us a generous commercial offer to figure out the issue - with the explicit caveat that she needed a complete and robust diagnosis in 14 days. We thought it was fair and we agreed.

Anxiety sets in. Have I made the biggest mistake in my career ?

Hours after the euphoria of an early client up-sell had faded, anxiety set in. Here I was, a media generalist, sales-oriented "suit" who had the gumption to question the veracity of a technical execution without even knowing properly what a cookie or a session was.

On a technical scale of 1 - 10, I was perhaps 2. I had a very poor idea of cookies, sessions, re-directs, parameters, nothing.

I had 14 days to figure out the solution to a vaguely articulated problem and I had no clue where to start…

The engines of learning roar....

The deadline loomed over us. I contacted a hot SEO firm in SFO turn by Caleb Whitmore, now the author of several books on GA and web analytics. . He offered to coach me and my team for 180 USD for intensive GA training over MSN chat. I signed them up for 10 hours.

I flew down a friend from India, Punit Parikh, an experienced web developer with keen persistence and problem solving skills. He stayed at my house and we spoke about nothing but GA analytics, cookies, UTM parameters, HTTP Referrers, 301 Redirects and 404 errors for the next several days..

While all this wondrous wisdom was permeating my brain, the solution was elusive. What was the EXACT PROBLEM and what exactly did we need to do to resolve the issue and fix GA?

The Solution dawns slowly but surely !

On the night of the 13th day the solution dawned on me, in a random way, just like Kekule dreamt about a snake seizing its own tail, which helped him visualise the structure of Benzene; or Schrodinger’s equation, which owed more to “mathematical intuition” than rigorous proof.

GA was set up perfectly well but was getting fed the wrong data.

The issue was the massive number of geographical domains that the client had, all of which had their own pages and were contributing significantly to overall traffic. These domains were linked to the master domain by meta-redirects. Because of this traffic that originated from Search Engines were being described as “Direct” as opposed to “Organic Search”. Because of this, direct traffic was being overstated and organic traffic was being understated. We needed to change the URL redirects from Meta redirects to 301 Redirects.

We did the necessary changes, some involving technical work in getting the client to edit their web configurator files (that's the beating heart of a website) followed by extremely anxious moments as we hoped the website would work and we should see the results that we hypothesised would show,

Victory is achieved !

Lo and behold, the traffic patterns changed overnight in the GA consoles. We cracked the problem.

Organic Traffic doubled, Direct traffic halved.

We gave the client 15% QOQ growth for 4 quarters, hitting all KPIS as well as our performance payouts - on 2X of the original baseline value that the client thought they were getting…..

Learnings: Trust your Gut and develop Gumption

The Gut: reacting sub-consciously to implicit patterns

Why did the data fail my sniff test ? In retrospect it was my Gut reacting to what was a skewed distribution of traffic sources. Like the famous Vitruvian man showing the perfect dimensions of the human form, traffic to a website also follows a pattern - or a split between direct, organic, and referrals. It was this "golden-ratio" that was being violated.

Gumption

Gumption is one of my favourite words, but lacking in most people. The paragraph below sums it up well.

"If you easily give up, and don't have a lot of confidence or smarts, then you are lacking in gumption. It takes gumption to get things done — especially difficult things. Someone who takes risks without being afraid has gumption. Having gumption is like having "chutzpah." We all could probably use more gumption. Like common sense, it isn't that common.."

I had the gumption to challenge the data and traverse an entirely new trajectory as a result. It gave me technical skillsets that improved my employability and gave me the courage to question data in my career going ahead.

And material for a riveting story that I enjoy recollecting again and again !