I recently came back from a conference full of eCommerce Managers, Digital Marketing Professionals, Chief Information Officers and Chief Technology Officers. To my delight there was plenty of messaging about increasing the speed and performance of websites as additional ways to increase conversions onsite.
Interestingly “caching” was mentioned as part of the presentation to solve website speed and performance challenges and that website owners should definitely use products like Varnish Cache. Implemented correctly, a cached website using software such as Varnish can:
Increase a website’s bandwidth and allow for additional headroom on the website
Improve the speed and response of websites to the http request from the users browser
Allow more traffic onsite and serve pages quicker
With this in mind, the outcome is to make websites faster and better the user experience. A well-researched and supported theory is that the by-product of a faster eCommerce website is that they make more money.
The theory is based on the fact that transactions increase and the overall value of the transaction increases when websites are fast.
That is all well and good and as an organisation with a history and passion of bettering the performance and optimising websites, we are strong advocates of tools such as Varnish Cache, but what other things should you consider when implementing Varnish.
Cost of implementation
What cost….Varnish Cache is open source and therefore it is free…isn’t it?
Albeit Varnish Cache is an open source product and used across many hundreds of thousands of websites, there are other costs to consider when Varnish has been recommended to you to optimise your site, such as:
Cost of infrastructure to host your Varnish instances
Cost to configure Varnish using professional services
Cost to change the Varnish configuration if your website application changes
Cost to change the web application to make sure Varnish works correctly
Cost and effort involved to problem solve Caching issues and bugs when Varnish is live in your production environment
As you are introducing another technology in your delivery stack, you’ve introduced another moving part. More moving parts means more to look after and more that can go wrong. These costs are somewhat obvious and directly attributable to a project of site optimisation using a Caching solution, but are there other costs to consider?
Indeed there are…..Varnish is a powerful tool and has immense power when implemented correctly, however, we’ve seen many many implementations of Varnish go horribly wrong. So what are other hidden on indirect costs of poorly implemented Varnish?
We’ve seen instances of poorly implemented Varnish being responsible of user sessions being shared. In one particular instance when items were added to cart, for User X, this was being cached and being served to User Y. This lead to a loss of trust on the site, bad publicity in social media and a reported loss of loss of business. This particular organisation saw a downturn in conversions for months until they managed to regain the trust that was broken. This situation came around because of poor configuration of Varnish and lack of testing.
Lack of effectiveness
As Varnish is so powerful, if your web application is not built in coordination with your Varnish instance, features can get turned off. This would mean that you could have an install of Varnish, running on servers and being managed that no effective outcome on your HTML…..do you know how effective your Varnish configuration is on your website? There are tools like VCLFiddle which can assist with your cache hit / miss rate, but with a lack out of the box metrics to show your how effective your caching setup is, do you understand the impact to website performance Varnish has for your particular site….ask the question?
Lack of SSL
The challenge of SSL is widely recognised as Varnish does not do SSL. There are solutions available and there are plenty of articles written in to achieve this. But as an upfront caching strategy as an aspect of your delivery stack, this needs to be considered.
I am often told “we are already running Varnish and we have that under control” but when I ask the questions…
How effective is your VCL config?
What metrics are you using to measure the impact of your Varnish install?
How does your VCL fit into your development process?
What Cache hit / miss ratio are you achieving?
Do you use Varnish for static assets, html, dynamic html, video etc?
…I find that website owners often don’t know the answers to these questions. So it begs the question…what are you paying for and what is your ROI on your Varnish install?
We built section.io to solve these challenges. With section.io you can:
Implement Varnish with no infrastructure costs or costs to deploy Varnish on your site
Receive metrics around the effectiveness of your Varnish install and its value to you
Configure and test Varnish much more effectively and blend Varnish into your overall website development process and the way in which it works with your application
Solve the challenge of SSL
Why not sign up for free now and get the benefits of Varnish and lower your True Costs of Ownership.