Analytics guru Avinash Kaushik shares advice on how to optimise your investment in analytics tools and staff.
Expenditure on analytics continues to rocket, as businesses work to get a grip on the rising volumes of data they have at their disposal. According to Gartner’s annual CMO Spend Survey, analytics investments is the single biggest area of spend for today's marketers, representing nearly a tenth of their budget.
But while expenditure may be on the rise, the concern is that investment is a knee-jerk reaction to fears about the growing volumes of data, and that businesses are throwing money at the problem rather than targeting their investment.
With so many analytics tools on the market – ranging from free to fairly expensive – it can be a confusing process. Which one is most appropriate for your business?
“I grew up in the traditional business intelligence world, and we were constantly on a quest for the single source of the truth – one tool that could master and do everything for the business,” he explains. “That paradigm doesn’t work on the web. It is futile to think that you can get all the answers from one tool on the web, that's just not how the world works because we live in an insanely complex world.
“The framework that I outlined in Web Analytics 2.0 said that as you mature you need to think of five different sources of data in order to make comprehensive decisions and that's: clickstream, outcomes, experimentation, worth of customer and competitive intelligence. These five sources of data answer four of the very important questions about your business - the what, how much, the why and the what else. Any company that is going to just pick one tool to measure success is not going to have spectacular success because they won't be able to answer all these four questions.”
Put this way, it would appear that analytics investment is an even more complex beast than we dared imagine. But while this may be tough news to take, there is a silver lining.
There are so many companies that have built such amazing tools that allow you to answer these questions for free.
“The good news is that you can actually answer pretty much all of these questions including a free tool in the marketplace,” emphasises Kaushik. “There are so many companies that have built such amazing tools that allow you to answer these questions for free. If you're a large size company you're absolutely welcome to pay for it - but you don't have to if you’re a small to medium-sized business because you have a world-class option for free in any one of these five categories.”
He continues: “The second piece of good news is you actually only need one tool to answer each question. So the thing that you don't want to do is to have three clickstream tools inside you company - that's stupid. You can just use one - pick the one you like, stick with it. You don't need 19 experimentation tools - that's stupid! Pick one and stick with it. You don't have to use 15 company intelligence tools, just use one.
“The new paradigm is that companies are going to have to have at least five different tools in order to get a comprehensive view of their business. The good news is that there is a free option available in pretty much every category and that they answer different questions. Don’t have multiple tools to answer the same questions. If you do, you’re setting yourself up for failure.”
What is the 10/90 rule of analytics investment?
But analytics investment isn’t just about tools – firms also need to have the staff resources to effectively use web data analytics. And according to research, businesses have a tendency to be under-resourced in this capacity.
The 2017 Online Measurement and Strategy Report by Econsultancy
has revealed that while analytics expenditure remains buoyant for technology, expenditure on internal staff is stalling. The research - based on a survey of 350 digital marketers and web analysts - found that 50% of respondents reported that staff budgets would remain the same, while 11% actually reported they would be cut.
This compares to 52% of respondents saying that budgets for analytics technology will increase - and only 3% suggesting that expenditure on tools could decrease.
The good news is that the number of organisations without a single dedicated in-house analytics expert continues to drop - falling from 21% in 2014 to 16% in 2017.
Analytics investment isn’t just about tools – firms also need to have the staff resources to effectively use web data analytics.
Kaushik believes that as a rule of thumb, businesses should invest considerably more in staff resources than in analytics tools.
“In my first job in analytics, when I was working at Intuit almost 10 years ago now, I created a 10/90 rule,” he explains. “The 10/90 rule says if you have $100 to invest in making good decisions on the internet then invest $10 in tools and consultants to implement the tool. Invest $90 in the people who will analyse the data because in the web that's the point of failure.
“We don't have enough money invested in smart people to analyse data and that's why I’m confident that companies that embrace the 10/90 rule will succeed because it gives them a robust set of tools to use and the people to go back and do some amazing things with that data.”
As a big proponent of segmentation, it’s no surprise that one of those “amazing things” is to slice and dice data. And while commonly used segments include the likes of direct, organic search, paid search, referred traffic, social and mobile, Kaushik has some special hints and tips to take segmentation to another level.
“I have a simple framework that is established for segmentation because you could go in a thousand different directions,” he says. “We should think of all our segments in three categories - acquisition, behaviour and outcomes. Take a plain piece of paper and write these three words on top. You should then think about what we are doing to attract people to our website. That becomes the acquisition category. Then, once they arrive at our website, what kind of behaviour do we want them to do? The last one is what kind of outcomes did we deliver for the business, and secondly for the consumer?
“So now you understand that all your segments fall into three brackets and that you must have segments in each of the three categories. If you do not do that you’re not looking at your business end-to-end.”
He continues: “Typical segments you might have for acquisition are: segment data by the types of marketing activity we're doing and how well represented we are in those. That could be earned media, owned media and paid media. So we could create a segment around Google search and Yahoo! display and MSN video ads - they're three segments you could immediately create. Or you could simply create segments by geography and say 'well we're trying to get traffic to come to us from all five countries in the UK, or all by every county, etc.’
“Behaviour segments are very simple - it's people who do certain behaviour so we could say well we want people to be loyal to us so we're going to create a segment for people who come to our website more than four times in a month or a week. And then you can segment and analyse - where do these people come from, what stories do they read, what products do they buy etc. And then you can optimise your existence or you could create a segment in the behaviour category around people who watch videos on our website.
“The last segment is outcomes - that's when you segment people who end up buying something verses the people who don't, or people who download a whitepaper or people who don't. People who indicated that they were fully satisfied with the website versus people who are not, so those are the outcome segments themselves. Those are just examples of the kind of segments you can create in order to have some really innovative analysis and understand what is happening in the business that will have a longer term impact.”