Data Driven Marketing Decisions

“Half the money I spend on advertising is wasted; the trouble is I don’t know which half.” – John Wanamaker

decision-makingThe above quote from John illustrates how difficult it was to track a response to an ad and find out which ad worked and which ad didn’t.

During the days of newspaper, TV and radio advertising, marketers went ahead with their gut feeling and sometimes they got it right and other times not so much. They had no other choice because there were no tools or technology to enable complex big data tracking and find out what worked and what didn’t.

Instincts and gut feeling can get you started in marketing but you need data to back it up if you want to move forward.

Bastiaan Zwanenburg in his article on medium says:

“Making decisions based on gut feeling has been very popular for years and years, and it seems that there still are many people who are really enthusiastic about this way of making decisions. However, the “digital revolution” gave us some new and very powerful tools to make better decisions, and you’re an idiot if you don’t use them.”

I cannot disagree with him.

At Exotel, being a start up with limited marketing budget we couldn’t afford to rely on our gut feeling – even if there is a high probability of getting it right. Going ahead with the instincts is an option only where the data is not available.

For example, when we are creating a new landing page with a new marketing message, there is no data to start with. So based on our understanding of our market and customers, we put up a marketing message based on our instincts. But once we start sending traffic to it, everything is tracked. (For more insights about how we drive traffic to our web pages, read our post on Inbound Marketing ).

A digital marketer’s dream would be watch the customer in person & look at their reactions through every section of the website, landing page or sales page… but since that wouldn’t be possible, we try to get as close to that as possible using the tools available to us.

A/B Split Testing

To test what words, images and other elements on the landing page works, we create two versions of the landing pages with different variations and then track which page performs better. Split testing is not new to digital marketing. Direct response marketers before the internet era have done split testing by sending 2 different variations of direct mail to 2 segments of subscribers and tracking the calls manually coming to 2 different phone numbers. Obviously only a few could do it because of the complexity involved in the manual tracking method.


Image Credit:

Split testing in the digital marketing world is much easier, faster and along with split testing you can track a lot more stuff. Also there are a lot more performance metrics that can be tracked like: Engagement (time spent on the landing page and click activity), conversion to lead (registration) and conversion to a customer (paid clients).

The most accurate data can be given by the last one which is conversion into a paid client.

For example, let us consider that there are two landing pages: version A and B.

  • Version A may send us 1000 clicks, 100 leads and 10 conversions.
  • Version B may send us 1200 clicks, 110 leads and 12 conversions.

In the end, it is quite obvious that version B performs better – not because it sent us 1,200 clicks but because it gave us 12 clients. 1,200 clicks is a vanity metric. I can send you 10,000 clicks for $10 but you would end up wasting $10.

Clicks and visitors don’t mean much because the end conversions are the ones which affect the revenue, the stability of the company, continued service to our clients and most importantly, our jobs!

Putting more weight on real metrics is important because sometimes version B can send only 900 clicks, 80 leads but may give 14 conversions. Vanity metrics can seem to lie some times. This happens when you refine your marketing message and disqualify people who are not going to be your clients. Lesser leads but more quality leads means more energy can be allocated to the fewer leads leading to more conversions in the end.

Such anomalies does not happen all the time, but it is important for you to note that end conversions matter more than the numbers in the beginning of the funnel.

BUT, you cannot always afford to have that kind of luxury – to track only the end conversions. Initially with low traffic numbers, the end conversions will be less. The sampling size will be too small and it will not help you make decisions. In such cases, little bit of data mixed with a little bit of gut feeling is the way to go!!!

So you may ask – “Deepak, What are you trying to say? To rely on data or gut feeling?” My answer is always: Begin with the gut feeling and wait for the data to prove you right, or wrong. When the data gives you the truth, your ability to get it right next time increases. Because you have now validated your instincts.

As you start nailing down traffic and conversions you can increase your numbers and sampling size. With large numbers, your conversions become more predictable and there will be less anomalies – that’s the law of large numbers .

As your conversions become more predictable, you can confidently derive results from your experiments and use the data to make decisions with conviction. That’s how you move from being a good digital marketer to a great digital marketer!

Now that I have convinced you that you should use data to make marketing decisions, let us look at the tools that we use @ Exotel to make data driven marketing decisions:

Visual Website Optimizer

VWO is one of our favorite tools for A/B split testing. Digital marketers can use their visual editor to make changes to the landing page and immediately start an experiment without waiting for the designer to make the changes.

Visual website optimizer mostly advertises itself as a split testing tool but they also have some amazing features like heat maps and click maps. Their eye tracking technology helps you determine where people are looking at your landing page. For example, here’s a heat map of our home page. A lot of people visit our home page to login and that’s where you can see that’s there is a lot of “heat”.


Google Analytics

Being a free tool and being a Google product, almost everyone uses Google analytics nowadays. Yes, we use Google analytics too but it is mostly limited to front end conversions. We track unique visitors, page views, referral sources, devices and a few more metrics like that.

For middle and back end conversions like conversion to leads, customers etc., we have observed that GA is not THAT accurate. Kissmetrics, a bit of custom coding and our own CRM gives better insights into back-end conversions. Also, GA only tells you how many people converted and not who converted.

Google Analytics also provides split testing functionality through its option called content experiments but we prefer VWO.


Kissmetrics is the holy grail of digital tracking. It requires a bit of tech for proper implementation but once it is deployed it can be an amazing source of data like no other. If a person visits a blog post on your website, then comes back after 2 days to check your pricing page and then registers on your website, you will be able to track this entire flow.

With Google analytics you would be able to track only the number of people who visited a particular blog page but not who. And you would have no idea if certain blog posts are sending you leads. Kissmetrics is a paid product and costs $150 per month but it is totally worth it – if you know how to use it properly.

Custom Coding

Apart from the tools mentioned above, we also have done a little bit of custom coding to help us track our traffic sources and conversions. With the help of UTM codes, we track the traffic sources and respective lead conversions and load the data into our CRM.

For example, if 100 leads came from Facebook, Adwords and LinkedIn each, after a few weeks we could determine the number of end conversions that each source has sent us. In our experience we have found out that certain traffic sources send us a lot of clicks and leads at low cost but doesn’t give us more paid clients. We have stopped wasting money in such sources. The money is diverted to the sources which actually send us paying clients, even if they send us lesser number of clicks and leads.

In the end, the only number that needs to go up in the presentation is the cost per customer acquisition!

Everything else is a vanity metric: clicks, likes, tweets, retweets, time spent on the site, followers and so on.


Hope you got a good idea about how to make data driven marketing decisions with various tools available today for digital marketers. As I have mentioned through out the article, you should choose data when you have an option but when you do not have the luxury of data, you need to make some decisions based on instincts.

But so far I have not talked about how to sharpen your gut feeling and increase the power of your instincts. You may ask: is that even possible? YES! It is! In my next post I will share with you the techniques and methods with which you can improve your gut feelings and have a higher probability of getting your instincts right. (But that’s limited to digital marketing though!) 🙂

AUTHOR: Deepak Kanakraju
No Comments

Leave a Comment

Your email address will not be published.