We, especially we the readers of this blog, often struggle with moving our organizations to be more data driven. The beautiful irony is that the bigger the organization the less likely it is to be data driven, inspite of large sums being spent on tools and applications.
It is possible to create truly data driven organizations , but for my "Guru" talk at eMetrics summit in Washington DC tried to tackle a much more solvable problem: Creating a data driven boss (or the boss's boss or the boss's boss's boss's boss or….).
My hypothesis is that not only is this a solvable problem, but that it is also a way for you to be in an environment where you can be challenged while adding value to your big / small organization.
Before we dive in I must say that it is my assumption that you actually want to do this and that you have the passion to fight the good fight. Glory and higher salary will be the obvious end rewards, but by themselves they are not motivation enough.
You have to have the passion to want to roll the big boulder uphill, here are your weapons………
Six Steps to Creating A Data Driven Boss:
# 1: Get Over Yourself
The absolute critical first step. You were hired because you bring skills that are unique. It is quite likely that you are smarter than everyone else when it comes to data skills . Hence you want to do amazing and awesome things and create a multi-dimensional statistical regression formula with fifteen variables that could predict the temperature of your website every second.
But your boss stubbornly wants a report that shows referring url's and trends in visits.
You are disappointed at how little value you are adding. Stalemate. Unhappiness. Lack of data driving anything.
You have to figure out how to talk to your boss and his peers at their, possibly less data smart, level. Remember it takes time for any organization to evolve and I find that lots of Analysts and Marketers let their ego's get in the way. "I can't believe I have to do all this silly stuff rather than….. " You get the idea.
Of course you do. :)
Learn how to communicate with your boss, or his boss. Give them what they want so that they will get on the evolutionary cycle.
Here's another benefit, your boss is aware of a lot more context about what it going on in the organization and in terms of strategy and focus and goals and everything else. That information is critical to your success, it will make sure that you have all the context and intelligence you need to ensure you are solving the right web analytics mysteries .
Solve for evolution, if things are not as instantly smart as you want then don't be discouraged, leave your feelings aside and communicate (really) and understand your boss's perspective. Then figure out how you can solve for him and not you. For now.
This is harder to do than you can imagine. But give it a try.
# 2: Embrace Incompleteness
Many of us come to the world of Web Analysis with experiences in traditional analytics where things can be counted to the last drop. On top of that we are classically trained to not to take risk and to only make decisions based on data we can swear on to be "accurate".
The problem is that we live in the most perfect imperfect medium in the world: the Web.
For now it is impossible to collect data perfectly. It is ugly, it is dirty, it is incomplete and no matter how hard you try it is not going to get perfect.
Yet we can't resist.
We obsess about silly things like cookie deletion rates. We go down rat holes of chasing down the last 5% of the deltas.
Not only is that effort not worth it, it is futile. Simple reason: You can't ever know what the total number of Page Views were on your site, much less every thing else.
And that's ok.
Now here is the silver lining in this dark sky. We live in the most data rich environments were we can find a ton of actionable insights.
Think about it. Taking our ads in Fortune Magazine is a completely faith based initiative. Based on the number of subscribers the magazine has you assume there will be an outcome. Or maybe we do some primary market research.
You can do much better than that on the web.
You can see exactly how many people got the "magazine", you can see how many of them read each "story", you can see the ads they were "exposed to", you can….. and that's just the basics. You can see were each person came from, what drove them to have a interaction with you, what the outcome was, did it result in a positive brand impression and so on and so forth.
Yet we obsess about the last 2% perfection, and we waste the opportunity to use 80% great stuff we have.
So make sure you are on first party cookies, tag as many campaigns as you can in a clean way, collect data and make decisions. Resist the temptation to be perfect, it is the enemy of good enough.
Embrace incompleteness and it will set you free. Both you and your boss.
# 3: Give 10% Extra
Organizations run on reports and so does your boss. They ask questions and you give spreadsheets. Then they ask for more and you automate the production of spreadsheet.
The other day someone was recommending that the only tool you should use for web analytics is excel! Then they proceeded to share best practices on how to write complex macros and reference cells and reduce the pain of creating spreadsheets.
Here is the problem with that: You were initially providing data, and now you are providing data without even having to look at it!
How is anyone going to find actionable insights?
Your boss is fifteen steps removed from data, you are closest to it. Yet now you have become a reporting squirrel (I was going to say monkey but that sounds rude).
As the person closest to the data is it not your job to Look at it? Look and understand and make recommend?
Make a conscious choice what job you want to be in: Reporting Squirrel or Analysis Ninja ?
[Bosses: When you pay people / consultants to do "data work" also make this conscious choice – are you paying the consultant to be a RS or a AN? Don't hire a Squirrel expect them to do a Ninja's job.]
You do that by actually looking at the data. Stare at the table. Go visit the website, click around, experience it, then go back to the data, connect the dots that only you can because you are the smartest person in the room.
Now give your boss 10% extra: Your insights that your boss did not ask for.
Make a recommendation. Tell them what is working. Tell them what is broken. Tell them that xxx or yyy is a better metric to answer the question.
You create a data driven boss by giving them something they can drive from data. Not by giving them spreadsheets or reports or only what they want.
At the end of each week ask yourself, did you give your required 10%? 10% extra is all it takes.
# 4: Become A Marketer
I find that great Analysts are not simply "data people". They are "customer people".
Yes they have all the qualities that we have talked about before (critical thinkers , curious, common sense, etc). But a delightful quality I have found is that are Marketers.
The reason is that different parts of the organizations care about different things but Marketing cares about the business with a very unique perspective.
If you want to change your boss and your company then you'll have to become a Marketer, someone with an understanding marketing principles, someone who can be a customer advocate / champion, someone who can evangelize the purpose of data in creating customer centric decisions.
Sales cares for selling, IT cares about keeping servers and sites up, Engineering cares about building things that can hopefully be monetized, Marketing cares about customers and, most of the time, they care about a longer term success and not simply meeting this month's quota.
No matter what organization you are a part of, you have to become a Marketer. Think like a Marketer and execute with that mindset.
Your job is to "market" your data in unique and innovative ways that solve for the customer. Get an understanding of marketing.
Take a course in marketing at the local university, read up on it (I subscribe to Seth's RSS feed), partner with the Marketers in your company and absorb.
Your boss will love you. Your career will soar.
# 5: Business In The Service Of Data. Not.
Lots of companies are data rich and tools "richer". In fact in many of them extensive data efforts to mine the logs and extract and xml and data warehouse it and mix it and merge it and clean it and build for scale and BI it and……..
No insights come out.
In the obsession about capturing, processing, storing, moving, shaking, baking data the core reason for doing all that is forgotten.
When the question is asked, rarely : What has all this complexity delivered for the company? The answer usually is: We have lots of reports and know up to the moment exactly what is going on via our blackberrys.
Classic sign of a ecosystem were the business exists to produce data to employ people to do all of the above.
The business does not exist to produce data. Doh!
Data should exist to serve the needs to the business and provide insights that can be actioned. Get that mindset if you want to change your management's mindset about how decisions should be made.
Do an inventory, ask around, how many decisions have been made based on data that can be traced directly to have added value to the bottom line revenue numbers? [Bosses: Great filter to apply for Consultants you hire as well, ask them that last question.]
It is important, nay critical, to constantly check yourself and ask if the business is really serving the needs of data or vice a versa.
So what does this mean?
When you undertake data projects apply this advice: Do small, deliver in a month, measure if it had an impact on the bottomline (even if small). If yes continue to invest more. If not dump it, time to do something new.
Traditional IT projects tend to be long multi year undertakings that used to deliver in the traditional worlds. That does not work on the web.
On the web things happen too fast, they get complex too fast, and every data project you undertake starts to decay almost immediately. Embrace speed and flexibility and 80% good enough. Implement, measure value, if yes move forward, if not kill.
We don't kill enough, we plan too far, we "implement" for too long, we don't think smart and move fast . Change.
Data driven decisions are not made when you spend 95% of your time in "getting" data rather than analyzing what little (or a lot) you have. You want a data driven boss? Spend 80% of your time analyzing data and producing insights.
# 6: Adopt A New Mindset, Expand Horizons: Web Analytics 2.0.
One final bonus tip. Expand the data you use to make decisions, move away from clickstream clickstream clickstream all the time.
ClickStream data is good at the What. We have tortured it to find insights. We have done the best we can with just knowing the clicks. It has worked ok, but it has not done spectacularly. Hence we have tried to take it to the next level by adding a bunch more clicks together and making it more complex. That is not stuck at all.
The other problem with a clickstream only strategy is that your bosses don't get it. There is lots of confusion, still, about Visits and Visitors and Unique Visitors and Sessions and…. So when you and I add / divide / subtract / multiply our clicks and page views and present analysis it does not have quite the impact we want, because at some level our bosses still don't get it.
That's quite ok.
Web Analytics 2.0 is your friend.
There is one important reason for that: with Web Analytics 2.0 you are talking your boss's language, that of Customer Voices and Competitors and, get it (!), Money!
Most web analysts focus on analysis with Omniture and WebTrends and Visual Sciences and HBX and Google Analytics and Coremetrics etc.
If you are one of those consider expanding your skills and experience to understanding and executing Surveys and Remote Usability Testing and A/B – Multivariate and Testing and competitive intelligence and so on.
Doing that will mean that you can represent the customer voice back to your bosses with qualitative data. It will mean that you can fight the HiPPO driven opinions with data beyond clicks. It will mean that you can kindle a small fighting fire in your boss by showing how your competitors are doing.
You boss might not care about clicks, but you bet your bottom she/he cares about customers and competitors.
As you create your own execution strategies do a Web Analytics 2.0 checklist. How many cylinders (above) are you firing on?
Can you believe that I did the above presentation in 12 minutes with eight slides? :) [If you were there share your feedback! Looooong twelve mins?]
Creating a data driven boss is not difficult. All it takes is some or most of the above six things. I wish you all the very best, may your days be brighter and your bosses more data driven!
Ok now it's your turn.
Please share your perspectives, critique, additions, subtractions, bouquets and brickbats via comments. Thank you.