There is a lot of discussion in the web analytics community on what percentage of your analytics budget should be spent on tools vs. people. However, the question I’m posing is what percentage of your company’s online revenue should be invested in analytics to begin with? (Aka, 50:50 of what?)
With free solutions like Google Analytics out there, I’m not surprised some companies initially baulk at the cost of an enterprise solution. (After all, it’s more than “free”.)
I thought it might help us all to understand what other companies are doing. Therefore, if you’re at liberty to divulge (keeping in mind this is anonymous – I’m not asking who you are, or what company you work for) would you answer the following two questions?
I will happily share the findings with the community once complete.
An example is below, under the poll.
Your company generates a total of $200MM in yearly revenue.
(Note: revenue, not profit.)
$100MM of that comes from the online channel.
You spend $1MM total on analytics
$300K spent on tools
$700K spent on people.
= 1% spent on analytics
0.3% on tools
0.7% spent on people.
eMetrics 2010 was a great experience. Lots of smart, interesting speakers, and great food for thought. Here are some of my favourite points, quotes, etc:
Jim Sterne: Social Media – Time to Rethink Your Marketing Metrics
“Yes, I like data, but I’m all about customer centricity.”
Purpose of social media is to:
Consider your influence
Check on your competition
Take action and
Larry Freed from ForeSee Results:Managing Forward: Moving from Measuring the Past to Managing the Future
Measure what matters most – your customers. Need to measure what constitutes success from their view, not yours.
The consumer is in charge. Because of the internet, it is easy to shop around amongst competitors. You no longer have to go from store to store but can check another store’s prices from a mobile device. Power therefore now rests with a cross-channel consumer.
Satisfaction drives conversion, loyalty, retention and word of mouth. Satisfy your customers and be fiscally responsible to survive and thrive in the years ahead.
Knowledge is power. Integration of metrics yields big dividends. Turn data into information, information into intelligence.
Social Media Metrics Framework Faceoff
Possible objectives of social:
Fostering a dialogue
Measuring social media is still new, and it is challenging to figure out the right KPIs. Sentiment analysis is not yet at a point where it can be automated, as sentiment comes from a reader’s reaction. (And note, people’s reactions are not all the same!)
Hardest thing about social is getting a VP/Director/etc to focus on why they want to be in social. What are the goals? Put channel last (e.g. don’t worry about whether to be on Facebook or Twitter) and think first about why you want to be in social at all.
PR and Analytics often using different tools to measure social media success, and often don’t mix or share information. However, ideally they should integrate and share knowledge.
Joe Megibow, Expedia
“I like to build stuff and blow sh*t up”
Expedia achieved success via data integration and uniting teams. Originally had Tealeaf analysts, datawarehouse analysis, BI analysts (etc) in different countries, with no global standards of definition and measurement. There were multiple (conflicting) data systems, and while decisions were made by numbers, accuracy is a question, and they weren’t being made on facts.
[No wonder the web analytics industry doesn’t yet have standard definitions and measurement. Expedia didn’t even have them within one company!]
Now a united team, with consistent definitions and integrated data systems.
Joe’s six lessions:
If you’re not working with your peers, you’re competing with them. Make sure people aren’t coming to different people for the same questions and having them compete against each other!
Learn from finance. CFO has “the truth”, and seasoned analysts who work with it. Find your source of truth.
Don’t just count, DO! Go from data collection to data action. Analysts can actually push projects through!
Sign up for results. Are you willing to bet your job, your team?
Manage expectations. If something will take a month, be clear about that. Business leaders just want plans and forecasts they can count on.
Start small, and communicate, communicate, communicate. Find little business wins, and proactively deliver. Earn the right to do more.
The moral of the story? Get stuff done. Take ownership, get results, get more ownership.
Vendor two-minute presentations
I think this is all that really has to be shared!
Adam Greco, Salesforce.com
Web data and CRM data is like peanut butter and chocolate – better together. Web analytics data is good. Integrated with other systems is what drives real value.
Add offline success metrics to online variables.
Segment web analytics data by CRM fields
Target web promotions using web and CRM data.
PS. Awesome success via integrating golf handicap from CRM system into ad promotion on site!
Add easy to understand value of a website visitor via scoring. (E.g. reaching Page X is worth 1 point, Page Y is 2 points, etc.) Every visitor therefore has a score/value of their visit.
To justify your analytics budget, you need to show them the money. Tie your success to revenue. It’s what CEOs understand.
Stephane Hamel: Measuring Your Organisation’s Web Analytics Maturity
Only way to success: Creativity in continuous improvement and attention to details. Generate innovative ideas and manifest them through to reality. Requires original thinking and producing. Analytics should use creativity. Analytics without creativity is just theory and pure mathematics.
Analytics = how a business arrives at an optimal and realistic decision, based on existing data.
Need people, process and technology.
Good thoughts from the session:
No data is enough if someone doesn’t want to believe. (Seth Godin)
Creativity is always constrained in some way. (Stephane’s daughter!)
Constraints make us push our limits (Peter Gabriel)
If data blocks your creativity, either your idea sucks, or you’re not being creative enough. (Jim Sterne)
Anyone can make the simple complicated. The challenge is to make the hard stuff easy. Analysis requires breaking a complex topic into smaller parts, to gain a better understanding of it.
Web analytics may be new, but we can (and should) learn from existing disciplines, rather than recreate the wheel.
The maturity model: Essential elements of effective processes which describes an evolutionary improvement from ad hoc, immature processes to disciplined, mature processes with improved quality and effectiveness.
Defined by level of maturity on a variety of axes. Aim is to develop but also remain even (e.g. no sense being mature in technology and having poorly developed analyst resources.) Growing one level of maturity is equal to one year of development.
Get down to earth
Fix your issues
Maintain balance (no point being well developed in one area, and low in another.)
Bob Page, eBay
eBay has an executive team that believes you can optimise the business with data. (And FYI, eBay has a ridiculous amount of data.)
They operate via distributed teams, organised datawarehouses, virtual data marts, but guided by common “North Star” metrics – those that are most important to the business.
Have developed an analytics community – like a Facebook for their many analysts. Allows information sharing, knowledge, and building relationships.
Structure of eBay analytics:
Dual hub & spoke model
Centralised technical team under the CTO
Centralised business analysts under the CFO
Distributed product analysts
The validation of groups’ findings helps keep a “separation of church and state” – keeps the businesses and teams honest when held accountable to another team.
Web analysts need to speak with a common vocabulary to and with finance.
Exploration and testing are core pillars of an analytics driven organisation.
If all of your tests succeed, you’re not pushing hard enough. You need to do silly things, and fail.
Ensighten Tag Management
Interesting new kid on the block regarding tag management. “Just because we’ve always done it that way, doesn’t mean it’s the right way.”
KPIs are your “Oh, sh*t!” metrics. If it doesn’t matter if they move up or down, that’s not a KPI. If they never change, that’s also not a KPI.
KPIs must also be actionable.
To make KPIs meaningful, they should be tied to people’s bonuses. Those are key for people.
There can (and should) be layers of KPIs: executives will have just a top-line few, middle will have more, analysts will have a lot. Being closer to the product and daily decisions will mean you should be looking at more detail.
Take the lead as an analyst in defining KPIs, and get buy-in. (If you ask for input, you may end up with 10,000 “Key” Performance Indicators.)
Agree on an expiration/review date, to make sure they get revisited from time to time.
Some think of KPIs as your dashboard, but John Lovett proposes thinking of them as your low fuel light, because you have to take action.
High level KPIs are great, but you must segment to make sure that they’re not hiding a lower-level trend. (E.g. One segment up, one down.)
If you’re not sure if a KPI is useful, or if anyone is looking at it? Go “metric radio silent” (or use fake data!!) to see if anyone notices!
Analysts, take the initiative, but collaborate. And don’t groan when a change in KPIs is needed …
Global Analytics at IBM
Analytics managers: Set a clear goal, give the team the right resources, and watch them soar. (Margaret Escobar from IBM)
IBM global web analytics team is a centralised and paid internal service. Internal clients contract with analytics at the beginning of the year as to the number of resources needed, the skill level, etc.
IBM focuses on:
People (need to communicate, especially in a global team)
Process (trying to establish “reusable assets” aka templates, as well as sharing methodology and focusing on documentation)
Trying to balance:
Consistency, with customisation
Demand, with overloading the team
Learning and exploring, but still getting the work done.
IBM product management: Need to embed metrics and analytics into your program develop, and decide what qualifies as success before you start.
“I received the oddest compliment today: ‘You make analytics and fun no longer seem mutually exclusive.’ But later (different person): ‘I’m not sure you do anything smart …’ ” – Lee Isensee
“Analysts take a complex topic, then breaks it up to gain a better understanding. Don’t throw data at people. Tell the story.” – Michelle Rutan
Question: “Is there a tool that will integrate data from various sources?” Answer: “You’re the tool.”
“Counting is not analytics. Seems obvious, but what did you do last week?” – @sutterbomb
“Analysts need to get better at the why, not the what.” Pat LaPointe, Michael Dunn session.
I’m headed home from eMetrics (in fact, I’m writing this from cruising altitude. Love in-flight wi-fi!) It was my first time at an eMetrics event, and I have to say – it not only met but exceeded my hopes and expectations. (And for the record, I was very excited to go, so my expectations were set high!)
I’m going to write a more in-depth post at a later date, but wanted to share a few thoughts. If these don’t suffice, feel free to check out the Twitter stream on the conference, which I will fully admit I dominated.
However, I wanted to share a few thoughts and favourite quotes of the day.
“I like to build stuff, and blow sh*t up.” (Joe Megibow, Expedia)
“KPIs are your ‘Oh, sh*t!’ metrics.” (Angie Brown) [Apparently it was word of the day …]
Kill unnecessary reports by trying to go “metrics radio silent” and see who notices (Lee Isensee, Unica)
The highlights for me were Joe Megibow, Stephane Hamel and Bob Page (with Adam Greco rounding out the list) and definitely the networking, conversations, and smart people I got to spend three days with. I feel very fortunate to have found such an amazing industry to be a part of, and can’t wait to get more involved via the Analysis Exchange, Web Analytics Association, and hopefully more conference presentations!
I think most conferences end up having a focus. This definitely seems the year of social and the holy grail of multi-channel analytics and complete view of the customer. (Note: we’re now getting demanding. It’s not enough to tie all online. We need traditional and POS and phone data – oh my!)
I also presented about analytics at Kelley Blue Book (the presentation is available here) which was a great first presentation experience. Someone very silly put me in the biiiiig ballroom, but happily, people showed up, and even had questions later!
I’ll share more once I have a chance to compile notes and tweets.
eMetrics DC 2010: “Marketing Metrics: The Publisher Perspective”
When your business model is advertising, your focus is different. Michele talks about the tools, techniques and unique KPI’s at Kelley Blue Book in this session about audience measurement, web analytics and data integration to forecast site traffic, on-site advertising inventory and revenue. Michele goes deep about statistical rigor, automated reporting and managing an internal advertising data warehouse.
So my blog was hacked yesterday sometime. With the help of the most awesomest person ever (my friend, Michael), it’s up and running with no lost posts. (My worst fear was that I had lost everything I had written.) I’ve tried to recreate/improve, but if you see anything that looks off, broken links, anything, please email me or comment to provide feedback. Thank you!