By now most of us realize site speed impacts our business metrics but many organizations large and small underestimate the impact. It’s well documented by now that as performance slows, bounce rates increase, pages per session decrease, conversion rates suffer, and increasingly organic ranking suffers.  As Google moves further to a mobile-first algorithm, the importance to SEO will only expand. Countless sites have acceptable desktop performance only to learn their responsive websites (in particular) suffer from egregious page performance. This is a side effect of QAing on desktop first, high bandwidth data connections, and general lack of awareness.

No one sets out to build a slow site yet in most organizations I’ve been a part of, that’s one of the main complaints, especially among the laymen in the organization. While engineers and product use the site day in, day out, they’re aware of the speed but also understand how difficult the task at hand is. Additionally, dev servers often have little to no relationship to production. In few organizations has performance stayed top of mind, and in none have they been able to bat 1000, especially without good processes and monitoring in place.

Product Measurements and Tools

If you’re not currently utilizing a product dashboard, well…get to it. If you are, making sure to incorporate performance metrics alone will drive decisions in the right direction. Page load is one of those areas where up and to the right isn’t a good thing so if you’re really cynical…use a negative Y axis :).

Here are examples of KPIs to monitor, bear in mind, we’re

  • Google Page Speed – https://developers.google.com/speed/pagespeed/
    • Score – 0-100 – As a product managers tool we use the score as an aggregate of the underlying components. By tracking this overtime we’re able to monitor where we are compared to our goals. Aim for 100% and chip away at your backlog.
    • Mobile Page Speed & Usability – Google allows this handy feature. When “Mobilegeddon” was approaching, I was part of an “all hands on deck” exercise to ensure we were prepared. It turned out to be a bit of a blip on the radar but by sharing these metrics with developers and teams you provide a simple benchmark by which to ensure you’re not caught on the wrong side of an algo update.
    • For Scoping: Once you’ve established your overall score, take a look through the best practices. These are the most common industry best practices and are handy for aligning your leads and engineers.
  • Synthetic and Real User Monitoring – Using tools like Catchpoint, New Relic, or your own homegrown tools, you can capture these two events.
    • Synthetic Monitoring – With this scheme, you have IPs spread throughout your market (around the country or around the world) and they fetch your webpage to gather the key metrics for your site.
      • Usage: This is best for benchmarking. By continually monitoring, you’re able to get visibility across the range of locations which keeping most variables constant. It’s the same selection or location, browser, etc.
      • Note: This is the best way to keep tabs on your competitors. By monitoring against the competition you provide solid performance benchmarks to motivate teams, understand changes, find new releases, and otherwise make business decisions.
      • Transaction and API monitoring – I break these out for one important reason. While initial page loads can be quick (or slow), transactions may be painful to customers. Especially if you see a funnel dropoff, you should consider scripting majopr site events to ensure they’re occurring in rational time periods. This is great for finding issues with databases and other load dependencies that you’d otherwise miss, especially on a cached site.
    • Real User Monitoring – By adding tags to your site you’re able to capture and aggregate the ACTUAL user performance. This adds a real depth to the above. Often you’ll realize that many users experience far worse than benchmark performance. This is especially true on mobile sites as connections are even more spotty.
    • Google Analytics – Google Analytics Page Speed tracking is extremely handy when enabled. The data gathered is less robust than a paid solution however it’s readily present in Google Analytics, allowing you to compare various user behaviors to page speed metrics.
  • Secondary KPIs – These KPIs are the indirect results of poor performance but allow us to combine performance date was business metrics. By quantifying business KPIs you will ensure you’re focused on the right issues and prioritizing correctly.
    • Bounce Rate – Bounce Rate will logically correlate with performance. Numerous studies by Google and others have shown the relationship between the two.
    • Conversion and Abandonment Rates – As I mentioned. If you’re tracking transactional scripts you’ll often find painful points in the process. By identifying these, especially under heavy traffic, you’ll see that users are less likely to abandon from slow sites.
    • Ranking Correlation – This is widely dependent on industry. Once you have industry benchmarks this

As we work through our site performance series I’ll continue to give insights and examples. I’m a firm believer that without proper KPIs you have no idea where you stand, what to prioritize, and how you’ll impact the end users.

Next time on the Ides Blog,  I’ll go into the various ways to look at performance, how to hedge with user experience and the SEO case for performance. 


Leave a Reply

Your email address will not be published. Required fields are marked *