#### Optimizing your Product

To run a business effectively and efficiently its important to know all the tools which can accelerate your growth. Metrics around your product are a way to keep a track of the growth and usability of your product. It helps focus on what’s working and kill whats not. Tracking and A/B Testing are two tools which come in handy while product development. While most people in the product development know what it is, there are few, especially the people who are new or starting out who can’t seem to wrap their head around this. To those of you, I hear you, I have been there, done that. But with time and experience I now know how it works and how to make it work.

To understand metrics and A/B testing, one has to first understand its need and why it is so crucial for any web product. Getting a customer to your site is expensive. Its called customer acquisition cost and once you have spent that money; you want the customer to convert i.e. make him/her pay some $$$ to you or do some activity like a facebook share or a tweet. There are two ways how one can optimize this.

First is the good old hunch. Visionaries seem to have an appetite for this. I unfortunately haven’t been blessed with a good gut. Therefore, I use the alternate process, the process of Metric Driven Development.

Metrics Driven Development is the process of collecting analytics around your product i.e. how many times was the product page visited, how much time was spend, how many links were clicked, etc. The best and the most popular tool for this has been Google Analytics(GA). You can signup for GA and start using it for free in seconds. As soon as you add GA to your site you can see (and in realtime also) how many people are visiting site, which pages are popular, etc. It gives an overall idea on how well you webpages are doing. This information is useful to track which pages are performing and which are not. Eg; if you are an ecommerce store and you see there is no traffic on your cart page after high traffic on your product pages, either your cart process/page is broken or the products on sale are really not interesting to the people who are coming to your site.

GA provides more than this. You can track events too. What are events? Anything, you want to track explicitly for the users coming to your site. Eg. You have a button clicking on which loads up a popup window with some information. You want to track how many times do users click that button.

Why? Because you care everything the user does on the site. To do is one has to add a couple of lines of JavaScript code (or HTML data-tags) and you can start seeing the data in your GA dashboard. More details on this can be found here.

Once you have your conversions going its time to optimize. This is where A/B Testing comes in. Most times, once a product design has been finalized, one gets sucked into a deep hole called d-e-t-a-i-l-s. It takes time to get the right design details. A/B Testing accelerates that, it is a technique where you create multiple versions of the design each with one of options you were confused with and then give it to random groups of people to find out which one is more liked. Most times, you have two versions of the same design hence the term A/B testing as you have a version A and another version B. When more than two versions are involved its called multivariate testing.

There are many tools for A/B Testing but none of them are as powerful yet simple as Visual Website Optimizer (VWO). With VWO you can start running A/B test with just a single line of code. Once you signup, you get a code to put on your site/page and after that you can run any number of tests by using its intuitive WYSIWYG interface. Its so simple that you will wonder if it will actually work.

What are the kind of tests one should run? A lot of people say changing the color gives a better conversion, I have never seen that happening. I would say getting a color scheme and following it uniformly is a better way to get the color right for your site’s design. ColorLovers.com is a good resource to find such color palettes. Changing in the font weight i.e. bold the text. Changing the content like using “Buy now” instead of “Pay now”. Using “Signup” instead of the word “Login” and so on give better results. The tests one has to run completely depends on your business and the kind of customers you are targeting. If you are new at this try checking what your competition is doing and then run A/B tests to see if it works for you. A/B tests on elements on a page help reach the local maximum for the page’s conversion. But in most cases, we need to optimize loops.

Loops or conversion funnels refer to the set of steps or pages a user has to go through so that our goal is complete. For example, for an eCommerce site; an user has to login, browse a product, add the product to his/her cart, checkout, add shipping info, make payment and finally see the confirmation screen. The way we analyze this is called funnel analysis; as the users funnel through the steps and after each step some users drop-off. The best tool I find to do this is MixPanel. Its a tool similar to the tracking features of Google Analytics but on steroids as it gives the data realtime and its tools are more focussed towards funnel analysis.

In funnel analysis, at each action of the user where one would expect the user to navigate to the next step one adds tracking. The login process is a good place to start a funnel, followed by login process success event, then browsing of items, navigating to a product page, and so on. MixPanel doesn’t guarantee the order of the steps and it should not concern you as its irrelevant as long as the user completes all the steps of the funnel. Once you have ran your experiment (I would recommend one complete week as not all the days of the week have similar traffic), its time to logon to MixPanel and see the funnel data. If you see a lot of people dropping off from a particular step of the loop, it means there is scope for improvement i.e. time to go back to A/B testing for that step and optimize.

I hope this helps people get started with Metrics and A/B tests on their products. If you have any queries do tweet me, I am @troysk704 and I will try and answer.

We are using all this and more to build; a lean startup; www.qreoh.com, a platform for independent designers.