Should a marketer worry about the varied streams of data portraying seemingly different stories? “I would think not,” says Seby Kallarakkal, the founder and CEO of the Bangalore-based Nabler Web Solutions Pvt Ltd (http://bit.ly/F4TSeby). Having data from other marketing streams would actually help a web analyst draw the right conclusions, he reasons during a recent interaction with Business Line.
“Let me give you an example. We were studying data for a website and found a big spike in traffic. And a few days later, the traffic went back to the old level. There was nothing on the website that could explain this sudden jump. But then we got some additional data – the dates on which the customer had run a TV add. The jump in traffic coincided with the relaying of the TV ad. And after a few days, the TV ad stopped. And gradually the traffic on the site also dropped.”
Something like this would have been hard to understand without the data on TV commercials, adds Seby. “So, I would think it helps to have data from multiple sources. It helps you build a more cohesive analysis on the traffic.”
Excerpts from the interview.
What’s your opinion on the effectiveness with which enterprises utilise their home page?
I do not want to answer as a designer. Let me answer that question from the perspective of being an analyst. Of course, the first thing to realise is that your home page is going to be seen only by perhaps half your visitors or even less. So if you have something to say that’s very important, do not mention that just on the home page.
Here are a few helpful hints based on our experience in analysing home pages:
a) Any kind of flash animation with ‘skip intro’ link is asking for trouble. Unless you are sure that you have a fabulous animation, do not do this.
b) Design the home page with some expectation. Have an expectation on what the user would click on. Then measure if that’s happening. If it’s not happening, change something. But do measure again.
c) When you measure, do segment. What repeat visitors do on a home page might be different from what new visitors do.
d) There are other useful segments – weekend visitors vs weekday visitors, visitors from different geographies, visitors at different times of the day. Their needs could be different and it’s worthwhile leveraging this difference.
Can web analytics be useful for analysing trends?
My answer would be, ‘Absolutely yes.’ One of the simplest ways of doing this would be to look at the keywords used by visitors to your website. Pull out the data for a year and you can see distinct trends that are happening within your industry. Of course, there are other factors like how popular you are with search engines. But from our experience, we have seen distinct trends based on keywords.
Once you are done looking at the data on your website, you can check Google Trends (www.google.com/trends) for similar trends. Google Trends will have data at a more macro level. If you want to see an example, try searching for ‘saas,’ ‘cloud computing.’ You can see how searches for saas are going down and how searches for cloud computing are going up. We will not debate on the difference between these two phrases. But Google Trends certainly shows how people refer to concepts and ideas, and how it changes over time.
In what ways can web analytics help a journalist?
I can think of a few ways in which a journalist can benefit from what web analytics could tell about a website.
a) Understanding the life cycle of an article. A journalist could get an understanding of when the popularity peaks and how long the article would take before it stabilises at a low traffic. You could also study those peaks in traffic happening after many months. How did those happen? Can you repeat it?
b) Understand how people came to the article. Did they come on the strength of the website? Or did they come directly to the article?
c) Look at keywords used by people who came to the article directly. Are you surprised? Was this the list you were expecting? Would you have rewritten the article if you had seen this list before?
d) Look at traffic from referring sites? Are you getting any significant traffic from any other websites or blogs? Were you expecting any? Would you do anything different knowing this information?
e) Use a tool like keyword tool from Google Adwords (https://adwords.google.co.in/select/KeywordToolExternal). The journalist might want to do a bit of research on what people are looking for before writing the article. Once you publish the article, compare the list you wrote for with the list that comes from actual data on your website.
f) Benchmark your article against your own and other journalists’ writing.
g) If you are promoting your article on various social media websites, track if your efforts are paying off. If it is, would you spend more time promoting them?
h) Look at how effective your cross linking has been. When you cross link your (or others’) articles from an article, do readers follow those links? Which ones did they follow and which ones they did not?
All these are questions that can be answered using web analytics data.
Would you like to outline a few low-hanging fruits for conversion?
Here are a few easy-to-implement ideas for increasing conversion on your website.
a) No prize for guessing this one. Identify a web analytics tool and install it. You can’t improve conversion rate if you don’t know what it is now. Once you have the web analytics tool, set it up correctly. Identify what you refer to as a conversion and set it up on the tool. You are ready to roll now.
b) Look at pages that are terrible at keeping the visitors. The web analytics tools would refer to these visits as bounced visits or short visits. It just means that visitors came to these pages and left the site without seeing any other page. Most of the times, this is bad news. Fix these pages. Rewrite them, add a message, do something. You have to stem the leak.
c) Look at conversions where the visitor has to go through a series of pages. In the web analytics context, this is referred to as a funnel. Wherever there is a funnel, typically there would be dropouts at each stage. You can work on reducing this percentage. You can try and reduce the number of steps or do something to stop people from abandoning your funnel. For example, if people are leaving on the payment page on the website, try to explain why your site is secure.
d) Learn testing. You can start with something as simple as A/B testing. It means you study two variations and choose the better converting one. You can do this manually and keep a tab on when and where you have maintained the two versions or use a tool. Again the tool from Google for this is free. It’s called Google optimiser (www.google.com/websiteoptimiser).
e) Look at your daily traffic and decide for yourself the threshold number you need for running your experiments. If you get just 10 visitors / day, no amount of optimising will lead to more conversion. Do a short burst campaign and get people to visit the website. Then start experimenting and measuring conversion. Pay-per-click campaigns are quite effective if you want a short burst of visitors.
What kind of questions should a CMO seek from social media analysis?
a) What are people talking about my brand? Is it positive or negative? What’s the percentage split?
b) What’s the volume of people talking about my brand? What’s the frequency? When are they posting?
c) Specifically, which features of my product are being appreciated by customers? Which features of the product are frustrating the customer?
d) Who are the influencers in the industry? What are they saying?
e) Which sites have conversation about my brand? Where do I need to step in and respond?
f) If there is something that I need to communicate about my brand, where do I make myself heard?
g) What’s the effectiveness of my social media efforts?
h) How am I doing vis-à-vis my competition?