The way forward for web analytics

With feet firmly on the ground, in Nageswara Rao park, the topic of discussion that Seby Kallarakkal and I engage in is the ‘future of web analytics.’ He begins by asserting that tools will become more intelligent. We have been seeing web analytics tools with more features, more customisation and more reports, elaborates Seby, the founder and CEO of Nabler Web Solutions Pvt Ltd, Bangalore (http://bit.ly/F4TSeby).

“Soon, we should be seeing web analytics tools moving towards more intelligence, giving analysis, conclusions, and perhaps even recommendations. If no one does it, we would be very tempted to build one.” Who can hold a ‘building’ task from someone with an IIT Kharagpur architecture degree, and a BITS Pilani MS in software systems, I remind myself…

There will be more tools in the market, even as the cost of developing a web analytics tool gets cheaper, he continues. “We are going to see web analytics tools developed by smaller companies. And these tools are not going to follow traditional reporting styles. They are going to experiment with different metrics and different presentation styles. They are also going to appeal to niche target audiences, or focus on niche areas of web reporting.”

Two other trends that Seby foresees are the integration with mainstream BI (business intelligence) tools, and an increase in data sources. While the former is already happening, we will continue to see more integration of web analytics data with the existing BI tools within the organisation, he explains.

As for handling more data sources, Seby cites as example the data from social media sites. Shouldn’t we then be talking about marketing analytics tool rather than of web analytics, he suggests. Our conversation continues over the email.

Excerpts from the interview

Is there too much data in web analytics?

Well, if you ask an analyst, there is never too much data! But, yes, there is a lot more data in web analytics than in other fields that I have seen. Imagine a site that gets a million visits per day. That’s a lot of data.

Another problem is the number of reports. Log into a standard web analytics tool and you will see more than 100 standard reports. If you are not clear on what you want, you can get lost pretty soon. My suggestion would be to have some questions in your head and then look at an analytics tool. That way, you would be focused on finding your numbers and will not get lost with rest of the data points.

Can analysis be automated?

That’s an interesting question. We have been exploring that for a while now. I believe that creation of the analytical framework can never be automated. A human being needs to do that. But after the framework is created and the decision rules written, quite a bit can be automated. We have written few like that, and I believe more will follow soon.

Any examples of customers benefiting from web analytics?

We do analysis for a B2B lead generation website. We have been working on this site for many years now and have moved everyone to make decisions based on data. And we have seen results in terms of increase in conversions.

Getting data that can be trusted is certainly a benefit. For example, we had started work for a customer and the data was messy. No one was using the data because they couldn’t trust the data. We got the customer finally to a point where the data could be trusted. So in this case, getting data that is trustworthy was a great benefit.

Ability to understand how things go viral is another great benefit. We were doing analysis for a beverage company and we noticed this big spike in social media about this brand. Digging deeper we realised that the spike happened due to a promotion run by this company in relation to the World AIDS Day. Understanding what causes ‘virality’ is one way to try and repeat it.

Should a CFO care about web analytics?

Web analytics gives a great framework for generating ROI (return on investment) for the money spent on digital marketing. If you look at the advertising pie, more and more money is going towards the digital marketing area. Having a sound web analytics setup can tell a CFO the return on all the money that is spent on digital marketing.

Would you like to trace the evolution of web analytics?

Well, I have to make this short. To begin with, there were the log files which log everything done by the visitors. In the earlier years of the Internet, web servers were rather vulnerable to the load. Neither the servers nor the bandwidth was very impressive. So web masters had to keep a close watch on what was happening on the website. So, they started using the log that was written by the web servers. It gave them information about every single hit to their server.

Soon, someone smart figured out that you can use this log file to count the number of visitors to your website. That was exciting. The marketing guys too came to know about this and they too loved this ability to see how many people are coming to their corporate websites. Companies started offering log analysers that would crunch log files and show clean reports to the marketing users.

However, log files were messy. They were large, difficult to handle and difficult to get. It used to fall under IT, and the marketing guys had a tough time getting ownership of the log files. Log files kept disappearing and along with them went the ability to look at historical data.

The next significant milestone solved this problem. And that was a small JavaScript code that you could embed in your webpage (page tagging). Soon, the marketing man had ownership of the web analytics data.

The discovery of getting data through JavaScript inspired a bunch of companies. Soon, there were many innovative products using page tagging. It was also now possible to have multiple web analytics tools on a single website without breaking your head.

I would think that the launching of Google Analytics marked another significant milestone. GA with its free offering was an instant hit. There were concerns about big brother knowing and big brother watching – what if Google knows what is happening on my website. But free is a very strong argument. One great thing that happened with GA was that availability of data became democratic. Almost all site owners now had access to a decent web analytics solution.

The next wave could be social media websites. The mainstream tools are yet to have any serious social media measurement capabilities. That’s perhaps going to change.

What are the challenges of doing social media analysis?

Too much qualitative data: Think about this – let’s say you have 10,000 tweets on twitter about your product. How do you even start to analyse these tweets? There are applications that could automatically tell you the sentiment. But if you want a better view, you will have to get deeper. That’s time consuming.

Unpredictability of where the data could originate from: Today, all social conversations about your product might be happening on a website. Tomorrow it could change to another website. Keeping track of multiple websites is a challenge.

Proliferation of tools: No one tool answers all your questions. So you have to use multiple tools and make them work together to get your data.

No known framework: There are different articles on various websites on how to measure social media effectiveness and how your product is doing on the social media. But the frameworks they are proposing are yet to mature.