The 10/20/70 rule for Achievable Web Analytics Success

In San Francisco during the “Guru Breakfast” event Rene Dechamps asked a question about the importance of process to web analytics. This is clearly something I believe to be tremendously important (quit my job, printed business cards, etc.) and Bryan Eisenberg commented that “web analytics was 10 percent technology, 20 percent people, and 70 percent process …”

Rene said he would post the video he took of this conversation soon, but suffice to say Avinash Kaushik, Jim Sterne and I all agreed with Bryan. Process is very important to web analytics, but the importance of process is often overlooked.

Recently a reporter got me thinking about these numbers, so I would like to formally propose an update to beloved guru Kaushik’s widely quoted 10/90 rule. I call it the 10/20/70 Rule for Achievable Web Analytics Success. Here is what it says …

Our Goal: Highest value from our investment in web analytics
Percent of time and effort spent on the selection and deployment of their technology platform: 10%
Percent of time and effort spent on the hiring and allocation of really smart people: 20%
Percent of time and effort spent on the process of actually “doing” web analytics, leveraging both technology and people: 70%
Bottom line for Achievable Success: It’s the process

The explicit recognition of the value of process resolves some of the issues people have with Kaushik’s original proposal, two of which include:

It is very difficult to spend $90 of every $100 on “intelligent resources/analysts” given the extreme dearth of available talent relative to the number of jobs currently open in the market today. Even Kaushik’s former organization (Intuit) is, to the best of my knowledge still looking for his replacement, several months later, ironically highlighting the difficulty of finding good talent.
Unless you’re gonna go the Google Analytics route (spending $0 on technology) and hire inexpensive resources to install the software (likely not one of the GAAC partners, although I’m not entirely sure what they charge) you’ll be hard pressed to spend $10 of every $100 on software license and implementation.

Now, I obviously agree with Avinash’s emphatic call to hire smart people. I’m a huge fan of dedicating resources to web analytics projects and have been since 2004 when JupiterResearch published my report Web Analytics: Spending, Staffing, and Vendor Selection. You need bright people to run your web analytics applications and to analyze data (although you may not need the people you think you need … more on that in another post at another time!)

But I think that the right way to frame the right approach to web analytics is not in terms of how you spend your valuable money, it’s how you spend your valuable time. So the 10/20/70 rule updates Kaushik’s rule by applying the appropriate emphasis squarely on the processes involved in “doing” web analytics.

Remember, you can always make more money, but it’s hard to make more time. Fortunately, some pretty bright people seem to agree with me.

Consider this: The technology involved is largely the same, especially at the level of need that most companies currently have from their web analytics solution. And while people are a good proxy for true process, in my experience too great of a dependence on people can cause two substantial problems:

If the people are not the right people, the organization may not realize there is a problem until a great deal of money has been spent and a great deal of time has been wasted
In my recent web analytics survey (results coming very soon!) we found that HALF of all respondents having web analytics experience had considered taking a new job in the last six months

So in the absence of process, some companies end up hiring unqualified people, hiring the wrong people, or hiring people who jump ship when the next best offer comes along. Certainly this is not the case with all companies, but until your organization has clear expectations about the goals for your investment in web analytics and how you plan to achieve those goals, technology and people will only get you so far.

I know, I know, I said to hire people and everything would be fine. It will be fine, but with process, you can be better than fine. You can make money. Piles of it.

So if you think your company is not following the 10/20/70 rule, here is my humble recommendation for you to consider:

Take whatever technology you have already deployed, until you’re good at web analytics the technology doesn’t really matter
Gather your key site stakeholders together
Ask them to share their experience and understanding of web analytics thus far
Document the gaps, looking for statements like “concerns about data accuracy”, “problems with data collection”, “not getting the right reports”, “reports are not actionable” and “concerns about how effectively we’re using web analytics tools”
Pick any typical site process such as launching a new campaign or deploying a new page or micro-site
Diagram the process you picked in step #5, highlighting decision points, tasks, and sub-tasks
Determine where measurement fits in the diagram you produce
Ask yourself if measurement, reporting, and/or analysis always happens in the places you’ve identified
If not, ask yourself if the lack of measurement, reporting, and/or analysis results in the stakeholder concerns discovered in step #4
If so, add measurement, reporting, and analysis to your diagram and make sure to follow the new diagram/process/checklist every time!

Still on the fence? Here are some questions for you to consider:

If you have spent $10 of every $100 on technology, are you successful in getting most of your questions answered?
If you answered “yes” to question #1, are you sure you’re asking the right questions?
If you’re trying to spend $90 of every $100 on people, how exactly is that going for you?
If you answered “great” to question #3, are you sure you’re not paying too much?
If you have great technology and great people, what is your web analytics ROI?
If you don’t know the answer to question #5, why not?
If you have dedicated analysts on your staff, what percentage of their time do they spend generating reports and attending meetings vs. producing analysis and managing experiments?
If you answered “too much reporting and meetings” to question #7, why do you think that is?
If you have thought about the process of doing web analytics, do you have a checklist or business process diagram for the core processes?
If you answered “huh?!” to question #9, call me.

Obviously my clear bias is for companies to invest in the process of doing web analytics. But professionally I have spent a great deal of time looking at this problem from all possible angles. And every time the answer is the same: The companies that invest their time refining how they actually “do” web analytics get more out of their efforts than companies who simply invest their money.

As always I welcome your comments and criticism.

Posted Sunday, June 3rd, 2007 | 28 responses | Add a Comment | Share, Save or Email

Steve

Aside: Gotta love timezones – Eric and Judah both publishing during my early evenings. Perfect!

I only disagree with one area in what you have here and what Avinash mentioned with his 10/90 rule. Mainly because it’s a problem for us.

Analytics is pretty much useless without the corresponding ability to change anything.

I appreciate it may be implied, but for some of us, it’s a killer. WA is emphatically not my main job, even so it takes months of subtle effort to effect even small positive changes out of the analysis I do provide.

I’ll start getting cynical again…. ;-)

eric

Steve: I’m heading back to my own timezone right now but I can vouch for Judah the “night owl!” Crazy huh?

Ironically you’re not disagreeing with me at all when you remind us that the “ability to change anything” is critical and fundamental. I said this at Emetrics (twice, so I must believe it):

“If you’re not running controlled experiments, you’re not doing web analytics!”

Controlled experimentation is a CRITICAL web analytics process, one that I outline in my white paper “The Web Analytics Business Process: Making the Case for a Process-Driven Approach to Web Site Measurement” (download here) But simply having people (i.e., 10/90) doesn’t create the mandate to run experiments and make changes. More often than not, hiring people in the absence of a clear PROCESS to make changes only aggravates the situation: You’ve spent money on technology and people, but aren’t seeing any positive ROI. Oops!

You are not alone, and that is the problem. IMVHO 10/20/70 is the solution.

Thanks for writing and I look forward to your cynicism!

daniel markus

Hi Eric,

Very interesting post. Food for thought, like always. What is hard for me to follow is the 70%. Defining and deploying the process seems a one time effort. Therefor I tend to stick with Avinash 10/90 rule.

Process is necessary, but how do you envision continuously spending 70% on the process?

You mention 70% of time and effort. But spending on process is not mutually exclusive from technology and people. Unless you mean hiring specialists to define and deploy the process.

When the process is in place, the analysts will spend the majority of their time on analysis and experiments. If the process is ill defined the analysts will spend time plugging holes with producing data and interpreting standard reports.

But when the process is in place you end up in the 10/90 situation again, where the smart people periodically review the process.

eric

Daniel: Perhaps what is confusing is the definition of “process” and I would encourage you to download my white paper on this subject from:

http://www.webanalyticsdemystified.com/wp

When I say “process” I mean the multitude of things every organization needs to accomplish to be successful with web analytics: assigning ownership of web analytics, defining business objectives, hiring people, determining an overall strategy for web analytics inside the organization, etc.

When you say “when the process is in place” I think perhaps you’re taking something for granted. Last week, in Europe, I only heard from a single company that had “the process in place” (Postbank) and I would offer that even they have struggles with the effective usage of web analytics elsewhere in their organization.

In my experience, often times the processes necessary to be successful with web analytics are only poorly defined. In these situations, as I said, the bright people you hire with Avinash’s $90 will struggle, be less effective internally than desired, or simply take a new job.

Again, I hope you’ll ready my white paper and thank you sincerely for your comment!

Rene

Hi Eric,
As we had the opportunity to discuss in Amsterdam last week-end ;-), I believe that your rule (70/20/10) is a declination of Avinash’s. Who’s creating and managing processes? People, bright people.

Rene

Ps I’ll publish soon the video with Bryan, Avinash and you speaking about the 70/20/10 rule at the last Emetrics.

eric

Rene: I respectfully disagree with your assertion that 10/20/70 is derivative of the 10/90 rule. In his original post, Mr. Kaushik cites:

“Required investment in “intelligent resources/analysts”: $ 90.”

The implication is that a company should spend $90 of every $100 on ** analysts ** who can run the applications. He goes on to say …

“2) Cancel the contract with your favorite expensive analytics vendor and take that $50k or $100k or $200k and: 1) Hire a smart analyst for between $50k to whatever maybe your areas great salary 2) Put the rest of the money in your pocket.”

… which reinforces the idea that ** an analyst is always the solution to the problem **. As you, Aurelie, and I discussed on the way to dinner Friday night, a surprising number of companies ** don’t actually need an analyst as badly as they need a “web analytics project manager” **, a position that few companies currently have.

I’m not saying that analysts aren’t required — far from it — but I don’t think that the majority of analysts are hired into situations where they are free and able to completely re-design how web analytics fits into normal business process. That’s why I say if you rely solely on bright people you might end with some very frustrated people; conversely, if you focus your efforts on creating the right process, the bright people will thrive (and Avinash’s sage guidance on spending will be fruitful!)

However, the point of 10/20/70 is that it’s not how you spend your ** money **, it’s how you spend your ** time **. This is a fundamental yet subtle difference between AK’s guidance and mine on the subject of “how to be successful with web analytics.”

As always, I’m willing to be wrong about this issue. I would love to hear from (more) of you who read my blog and who are currently in “analysts” roles in your organization. The questions I’d love to discuss are “Does your company allow you to be truly effective?” and “How much web analytics process have you been successful in implementing in your organization?”

Thanks again for your comment, Rene, and I look forward to seeing the video from the WAG event.

Judah

The challenge with the 90/10 guideline is when you spend major bucks on a tool. Say you spend a cool $1M US. You then spend $9M on people? At 50k a pop that’s 180 people (and no web analysts work that cheap). Say you spend $100k for the tool, that’s $900k on people. At $50k a pop that’s 18 people. Good luck getting the expense allocation or finding that number of resources, especially if the business is “new” to web analytics.

What about when you spend zero on a tool, doesn’t that mean you spend zero on the people or 100% of the cost, which is nothing. In this case how would one derive the 10% to frame the rule?

So the idea of segmenting, deriving, and/or deconstructing the 90/10 rule to apply to different classes of businesses and across different existentialisms (like time) makes sense.

Perhaps a 10/20/30/40 time-based rule? 10% of time selecting and extending the tool, 20% managing, 30% on analysis (part of process), and 40% on controlled experimentation and outcomes measurement (part of the process)? Just riffing right now. I sense a forthcoming blog entry…

eric

Judah: Good points about the spending model … I know Avinash was just trying to provide good guidance but it’s not easy to apply those numbers to any but the most simple models. That’s why I chose to focus 10/20/70 on ** time ** not money.

I mean, you can make more money, right?

In terms of your 10/20/30/40 … analysis and controlled experimentation (and reporting, and tagging, and data validation, and, …) are all part of the PROCESS which is why I think Bryan was so right when he said “70 percent is process.” Again and again it is the processes that are so important … as I’m sure your employer can appreciate — if they lose you to some cool new startup consulting firm, they sure don’t want to lose all the hard work and effort you’ve done (and they’ve paid for …)

I don’t know about another blog entry (I’m so backlogged as it is!) but let’s plan on drinks when I’m in Beantown week after next and hash out the solution over several Sam Adams.

As usual, thanks for commenting!

Rene

Hi Eric,

I understand your position and I should have selected my words when commenting (sorry but english is my third language). What I meant was more that it is an evolution of Avinash’s rule. As Judah says, the 10/90 rule dosen’t make sense in all the situations. Your rule seems more logical and indeed you speak about time/effort while Avinash speaks about money. From my perspective Avinash’s rule has helped to show the market that products weren’t the objective neither the silver bullets; even if the rule is not rocket science, it has helped the market to go a step further and put the focus on what happens once you’ve got a WA tool. Your rule goes a step further and brings processes to the arena, thus a more structured framework.

Hope that this makes more sense ;-)

Cheers from Paris,

Rene

eric

Rene: I should be so lucky to have a third language ;-) Your comment makes sense, and you’re correct. Avinash’s 10/90 rule reiterated the guidance that many people had been giving for years — JupiterResearch, Forrester, Jim Sterne, Jim Novo, Bryan Eisenberg, many others — that it was not the tools but how the tools were used.

10/20/70 builds on that but again is a summary of what some pretty smart people have been saying for quite some time. And the message is still the same — it’s not the tools, but rather the processes through which bright people use the tools to create success in the business.

Enjoy Paris and please give my regards to Aurelie!

Web Analytics: Four Gurus For You (4G4U) - WAA (5min video preview) « Web Analytics .be Blog

[...] P.S. I haven’t selected this excerpt because it was my question, well in part only ; but also because it explains the 10/20/70 rule that Eric has explained in details recently in his Blog. [...]

Web Analytics Demystified » Blog Archive » Video from Jeremiah Owyang and the WAW Guru breakfast

[...] More recently I wrote a post on the 10/20/70 Rule for Achievable Web Analytics Success in which I outlined the importance of process to web analytics. A number of folks have since commented on the post but Rene Dechamps from OX2 was kind enough to post a video from the conversation that got me thinking about 10/20/70 (thanks Rene!) [...]

Web Analytics Demystified » Blog Archive » The Problem with Free Analytics

[...] This was my position in 2004 when I published Web Analytics: Staffing, Spending, and Vendor Selection while at JupiterResearch–that people are as-or-more important to your ability to “do” web analytics than the technology you select. My position today is slightly revised, given my views on the importance of process to web analytics, but until you’re doing moderately complicated things with your web analytics package, based on clearly defined business objectives, the technology you deploy is not the limiting factor: you are. [...]

4 Simple Ways to Improve Your Web Analysis | Digital Alex | Interactive Marketing Strategies, Web Analytics, Conversion and More | Alex L. Cohen | Alex Cohen

[...] Obviously, there’s no substitute for a talented web analyst (that’s where you should be spending 90% of your budget) or a web analytics process. I’ve also found that context is a crucial element. Here are 4 simple ways to improve your web analysis. [...]

Webbanalytiker.nu » Web Analytics Tuesday

[...] Avinash Kaushik har sin 10/90 regel som säger 10% teknik och 90% smarta människor. Eric har sin 10/20/70 regel som säger 10% av tid och kraft på att välja teknisk plattform, 20% av tid och kraft att anställa smarta människor och 70% av tid och kraft att skapa processer och att verkligen utföra webbanalys. [...]

Should You Only Hire Web Analysts With Experience? | Digital Alex | Interactive Marketing Strategies, Web Analytics, Conversion and More | Alex L. Cohen | Alex Cohen

[...] Process: Let’s Not Reinvent The Wheel – It’s intimidating enough to start in a whole new field, but if your company/department/guy in a cube is run completely ad hoc without standards then it’s going to be a frustrating learning curve. Trust me, I’ve been in the situation where I had to create something from the ground up–3 different times. It’s a very educational experience, but it’s also a lot to expect of a new hire. Just consider how much infrastructure and process you have and whether you’re willing to give your new web analyst time to build any. [...]

Alanat Coop News » Web Marketing and Analytics: Process, Talent & Tools

[...] revised http://www.kaushik.net/avinash/2007/04/how-should-web-analysts-spend-their-day.html to http://blog.webanalyticsdemystified.com/weblog/2007/06/the-102070-rule-for-achievable-web-analytics-…. I agree with both that the key to value and success in web analytics is actually taking action on [...]

Cossacks Breaking News » Web Marketing and Analytics: Process, Talent & Tools

Web Marketing and Analytics: Process, Talent & Tools

Futures News » Web Marketing and Analytics: Process, Talent & Tools

Business Daily News » Web Marketing and Analytics: Process, Talent & Tools

Website Analysis Method 101

[...] Much blog ink has been spilled over the topic of web analytics process as a way to operationalize measurement (here, here, here and again here). [...]

Web Analytics Demystified » Blog Archive » What is your web analytics communication strategy?

[...] sure you know, this piece is far from a no-brainer — whether you subscribe to 10/90, 10/20/70, or some other percentage-wise distribution of effort, I think we can all agree that people are [...]

Web Analytics Demystified » Blog Archive » How Yahoo! buying IndexTools changes Web Analytics

[...] suddenly companies wanting to invest in web analytics will be far more likely to take advantage of the 10/20/70 rule for web analytics success I described over a year ago, focusing their efforts on people and process and worrying less about [...]

Yahoo! buys IndexTools! « The Analytics Guru

Web Analytics Demystified » Blog Archive » Great jobs at Intuit; Our job board is now GLOBAL

[...] former home of web analytics super-star Avinash Kaushik and as a company that clearly understands the 10/20/70 rule for achievable web analytics success. While Avinash has moved on to Google, bright folks like my friend Dylan Lewis (who will be leading [...]

Marko Hurst » Blog Archive » Stop Treating Your Visitors Like Dogs

[...] The 10/20/70 rule for Achievable Web Analytics Success [...]

Google Analytics räcker för de flesta företag | Webbanalys

[...] % av pengarna ska läggas på analytiker och inte verktyg, eller om du tycker att Eric T. Petersons 10/20/70-regel är bättre är det tydligt att det mesta av budgeten bör spenderas på det faktiska [...]

Add a Comment