In a previous post, I reflected on the behavioral aspects of web performance work. As a follow-up, I'd like to provide a little more detail about how to tackle performance issues under certain scenarios. If you're having trouble integrating performance work into your day-to-day tasks, this is for you.
According to the Project Management Body of Knowledge, all projects fit within a larger system that is intended to provide value at an organizational level. Providing value is the key to making it work, but each organization will have a different level of sophistication with how they view performance. With that in mind, you can go down the road of discovery to learn about how a company's web presence fits in with their business and specific performance issues.
How does a company see its web presence?
Years ago, I read David Bryant Copeland's book, Senior Software Engineer, and one of the insights there was not writing code before you have a full understanding of the problem. The same thing applies to performance optimization. Not only do you need to understand the technical aspects of what is wrong, but you also need to grasp what value a company derives from their web presence before you can have a plan for improvement.
In some cases, it might be necessary to get background on the historical decisions that led to the present situation such as the objectives that were part of the current design, identifying the purpose and owners of the third-party scripts, and how the business uses a site as part of their overall strategy. You might end up talking to ad ops, sales, editorial, design, product developers, and project management to get the information you need.
As you're talking to people, tell them what you're trying to accomplish.
If a lot of that is known due to your work history with the company, simply fill in the gaps of any information that might inform the performance strategy. There is never a shortage of work to be done, but ideally, it needs to be prioritized by how you can show value to your company so you can continue to do work that has impact.
Performance in different scenarios
As far as prioritization goes, I will often work on tasks that have a sweeping impact across the site first. There are a few different scenarios that might help you decide what to work on.
It's important to note that your scenario might change given different factors, such as a greater awareness of performance impact based on prior successes, a new leadership team, shifting business priorities, and many other factors.
Scenario 1: known correlation of performance and KPI with cultural buy-in
If you know which performance metric has the greatest impact on KPIs and have a company culture that is supportive of performance work, you have the ideal scenario.
You should work on the item with the greatest correlation because that will provide the most value to the business. If your company is supportive of performance work, they are likely patient as well. As long as you communicate progress, you probably have some time to plan big, sweeping changes that have a serious impact on the bottom line.
Scenario 2: unknown correlation but cultural buy-in
If you haven't established the business to performance metric relationship yet, but you have cultural buy-in for performance work, I suggest starting with the Core Web Vitals metric that has the lowest percentage in the good category.
You'll be able to make changes and demonstrate progress as you chip away at the issues. The more traction you make, the more likely you are to get added time to address performance issues, increased visibility to your work, or approval for tools or resources that can help you tie the performance work to business metrics that will be critical in making the next leap forward.
Be vocal and share your wins often. The marketing really helps, and you might start to get the attention of interdepartmental staff who can be allies for performance.
Scenario 3: unknown correlation with no cultural buy-in
If you fit into this situation, it can feel lonely, but it's also really easy to improve things. If you're in this category, it's likely that your company had neglected performance, and anything you do will be an improvement. This is where I started with my performance journey at my current company, and I'm proud to say that this is no longer the case.
It's all about showing progress as fast as you can. Get the low-hanging fruit, and start bringing things back under control.
For me, that was addressing Cumulative Layout Shift because the p75 was in the moderate category, but there were only a few UI elements that contributed significantly to the layout shifts. Reserving space for ads and hero images was one of the first things I fixed. Within a few days, I was able to say that we had a metric that was important to my company for SEO reasons that was good at p75 when it hadn't been previously.
There was a great value-to-effort ratio there, and that is just what I needed to get some traction for performance work at my company.
Scenario 4: known correlation with no organization buy-in
If your organization knows the specific performance metrics that impact the business performance but yet doesn't do anything about it, that's indicative of a company who has different priorities. If this is the case, it's a tough situation. Fortunately, it seems to be very rare.
If you find yourself in this situation, you can either work on performance between other work or bake performance into your assigned work. You'll be limited to things that don't have large scope because you will not have much time to dedicate to it.
In all cases, the goal is to improve the performance of the sites and ideally work toward transforming into a situation where there is a known correlation between performance and business metrics. If you're already in the ideal situation, you want to stay there and periodically look for other correlations to continue to guide your performance work. Bringing the value will help make this happen.