THANK YOU FOR SUBSCRIBING
By Khilan Haria, Vice President, Head of Product Management, Treebo Hotels
Entrepreneurs in the commerce space are always keen on optimizing conversion rates to drive maximum returns. Driving new traffic is more expensive as compared to maximizing conversion rate for the existing traffic that’s already coming to the website. Moreover, focusing on retention could be an added advantage as acquisition continues to grow more and more expensive. Sharing a few key aspects that every Product and Engineering Leader should focus on their website, even in the early days, to drive higher conversion rate with the existing traffic.
Fast: Performance as a Feature
I would recommend focusing on performance right from the first version of the website. Depending on the Go-to- Market obligations and the expected scale choose either a Responsive or an Adaptive approach. If time and resource constraints permit, adaptive route should be preferred as having a separate mobile site allows for an optimized user-experience on mobile. Moreover new age approaches such as Progressive Web App (PWA) and service workers are good enablers in high speed and good load times. Getting a caching solution in the form of a Content Delivery Network (CDN) for static content in also essential in the first version. Moreover, picking an off the shelf solution to optimize image size depending on the user device and bandwidth could lead to amazing performance in the first version and this is again a low-effort activity to deliver for the returns it could generate. I have observed a close correlation between performance and conversion rate - especially for mobile sites and developing geographies. Research also suggests for every 1 second drop in load times, conversion could go up by 5-10 percent. Apart from conversion rate, performance is becoming more and more important for Search Engine Optimization (SEO) as well.
Accelerate Mobile Pages (AMP)—For mobile page speeds
With page load times becoming more and more important, we have seen newer innovations such as Accelerated Mobile Pages (AMP) from Google to drive better speed on mobile sites. Conceptually AMP takes us back to the older era of web development with pre-compiled static pages along with the added advantage of these pages being cached on one of the most powerful CDNs around - Google CDN. It’s important to focus on AMP for commerce websites to both get to best-in-class page load times and for Search Engine Optimization. AMP results show a thunderbolt symbol next to the search result and these truly deliver the best-in-class performance. Google has started taking this into factor for organic search results and it could become more important in the future. Word of caution: Integrating PWA with AMP is not trivial and needs to be done carefully to avoid any bad experience when a user switches over from the AMP version to the PWA version on click of any action from the AMP site.
Adopting a data sciences lens to conversion optimization could help understand the importance of several factors affecting conversion rate and get us to best-in-class conversion rate
Solving for our users and answering the basic questions that’s running in their minds as part of the purchase decision will help increase conversion rate. Imagine if a hotel booking website does not mention whether the price is for 1 night or the entire itinerary on the search results page. It will potentially become an added question in the users mind and solving for this alone could help uptick conversion. Another common user question for any commerce website is whether a particular price includes or excludes taxes. Solving for some of these basics could go a long way in reducing user friction. Reducing the number of steps in completing the purchase could also help in bettering conversion rates. From a user-experience point of view, having a clear primary CTA and using colors well is equally important. On top of this a clear hierarchy of fonts helps in driving user focus on the most essential area. Heatmap is a good first step in measuring the some of these aspects of user experience.
What can’t be measured, can’t be optimized
It’s important to measure conversion rate and understand how it’s varying across different traffic sources and at different steps of the user journey. First and foremost tool that’s essential to understand this is Google Analytics. For additional insights, clickstream analytics tools such as Amplitude or Mixpanel are very helpful. These tools help understand how a user or a cohort of user is converting at each step of the funnel. Video recording tools such as Inspectlet or Hotjar help in understanding the plausible reasons behind drop-offs at specific steps in the user-journey and build hypothesis for experimentation. Powered with these insights, a Product Leader can identify and iterate on improving conversion rate. Some of these insights could seem counter-intuitive and even the most obvious iteration could backfire. Hence taking an A/B testing approach is the preferred route to experimentation and new launches. Practical tip: Rather than integrating all of these tools separately, one-at-a-time, using a data aggregation platform or an analytics hub such as Segment could result in a faster time-to-market. There are a number of other benefits in using Segment such as its flexibility to switch tools without development effort, using numerous other concepts such as marketing automation, browser push, in-website chat and integrating Advertising platforms to further optimize conversion rate as well as attract new traffic.
Advanced Conversion Optimization
On a closing note, I would like to leave with the thought of using advanced data analytics to find correlation and establish causality between higher conversion rate and the various factors involved. Adopting a data sciences lens to conversion optimization could help understand the importance of several factors affecting conversion rate and get us to best-in-class conversion rate. We are already seeing applications of machine learning in not just understanding the factors, but also in optimizing some of these in real-time.