How we helped NASDAQ's performance rally by 114%.
Client
Nasdaq
Expertise
Performance
Engineering
Engineering
Deliverables
Web
The second largest stock exchange in the world, the Nasdaq website is visited by many millions of users everyday. Performance becomes critical at such scale, where even a single second of delay can cause the business to incur huge losses.
We started working with Nasdaq in early 2020 to help improve web performance, and so far, we've pushed the score up by 114%. The aim is to take it higher still and to equip the Nasdaq teams with the tools required to sustain it.
We started working with Nasdaq in early 2020 to help improve web performance, and so far, we've pushed the score up by 114%. The aim is to take it higher still and to equip the Nasdaq teams with the tools required to sustain it.
The problem -
Poor performance just got expensive.
The oldest electronic exchange in the world, Nasdaq had been struggling with a slow, bloated web platform for years. The site needed to be sped up for sure, but the problem never seemed large enough to be immediately prioritised. Until early 2020, when Google decided to take web performance into account in its page ranking algorithms, turning Nasdaq's performance score into an incredibly critical business issue.
Placed firmly at the top of Google's search results, Nasdaq sees roughly
7.5 million
unique users every month.
Losing even a small percentage of that user base to a competitor would mean a steep drop in revenue — a price that the company was not prepared to pay.
What we did -
Taming the beast.
We specialise in chasing performance at scale. With Nasdaq, the problems were especially complex. The web platform is built and maintained by teams around the world. We needed to wrap our heads around several years' worth of code, collaborate with multiple tech and design teams, and of course, never ship code to the world's second-largest stock exchange before the markets close!
We spent our first few weeks looking under the hood — analysing the codebase, evaluating the architecture, and getting to know the devs. Understanding the inner workings of both the tech and the teams at Nasdaq was critical to mapping out our strategy. Before long, we had a roadmap spanning several months, collaborating with the Nasdaq teams to improve web performance — one metric at a time.
We spent our first few weeks looking under the hood — analysing the codebase, evaluating the architecture, and getting to know the devs. Understanding the inner workings of both the tech and the teams at Nasdaq was critical to mapping out our strategy. Before long, we had a roadmap spanning several months, collaborating with the Nasdaq teams to improve web performance — one metric at a time.
01
Nipping it in the bud.
How does a platform of this scale fall into disrepair? Quite easily, to be honest. Nasdaq has vendors across the world contributing in various ways to the site, and it's hard for large, distributed teams to stay in sync. Each team focuses on its own set of tasks, unaware of their impact on the site as a whole.
Performance needs to be a priority, not an afterthought.
Seeking to build a performance culture at Nasdaq, we conducted training sessions for the dev and marketing teams, outlining a few basic guidelines and best practices for building and maintaining a highly performant site. We created checklists for the engineers and set up an LHCI server to help them keep an eye on performance scores. Interventions like these help us tackle the issue from the ground up, and are key to ensuring that a product stays performant long after we hand it back to our clients.
02
Ironing out the UX.
A good website feels smooth. It loads quickly, doesn't jump around, and behaves as you'd expect it to. Incidentally, a lot of the little things that go into creating that buttery experience directly affect performance. If you've ever visited a webpage where things shift around as they load, you'll know how annoying it feels. That's Content Layout Shift — a metric that devs specifically target to improve performance.
With Nasdaq, we were able to bring the CLS of key pages down by up to 73.91%, and reduce it to 0 for many components. That's a definite UX overhaul, but it has an equally large impact on performance.
Fix the UX first. It'll keep your users happy, and your product lean.
03
Targeting the Core Web Vitals.
Core Web Vitals are a list of metrics used to measure a site's performance. Our initial forays into Nasdaq's codebase helped us identify several opportunities to improve performance by targeting these metrics individually. We used PageSpeed Insights, web.dev/measure, WebPageTest, and our own internal tools to analyse potential improvement areas on the site. The data helped us prioritise high-impact tasks that would translate into immediate improvements in Nasdaq's Core Web Vitals.
Most of our work here can be broken down into 3 categories
Component level improvements, including removal of unnecessary dependencies and optimisation.
Page-level optimisations, which mainly consisted of changes that will directly impact metrics like CLS and LCP.
Improvements in the bundling and delivery of pages to the browser.
We started with the high-impact metrics and have been working our way down the list. All of our code ships to production at night, to make sure that there is no danger of downtime while the markets are open. We monitor the pages and components we ship these changes to constantly. Every update is tracked closely, documented in detailed reports, and has a tangible, measurable impact.
This is what optimising for performance looks like at scale — many little steps that add up to one giant leap.
The impact -
Up by a whopping 114%
The Nasdaq site had a performance score of 35% when we started work on it. The changes we've already made have had a huge impact on that metric, which now stands at 75%. The journey's not over yet. Our aim is to push that score up to over 85%. That last bit is the trickiest, but we know exactly what's needed to get us there.
75%
Performance Score
Up from 35%, and our target is to get to 85%.
96%
Improvement to CLS
Down from 0.25% to 0.01%
50%
Reduction in LCP time
Saving multiple seconds during page load
The road ahead -
Attaining peak performance, and planning a new future.
Fixing Nasdaq's frontend gave us huge performance gains, but that's just the start. So many aspects of web infrastructure have an impact on performance — our aim is to focus on each of them, one at a time. For now, API caching strategies and the way region specific routes are handled are under the scanner.
Performance, once gained, must also be sustained. Now that we have all of Nasdaq's teams on the same page regarding this, we are constantly creating systems to enable them to monitor the impact of deployments within the QA environment itself.
Zooming out, we are also working on some big picture changes. Technologies evolve constantly, often leaping ahead light years in a matter of months. Nasdaq would definitely benefit from embracing some of these newer approaches, and a complete restructure on the web frontend is being considered. This needs to happen without business being disrupted, though, and we are currently working on a strategy to conduct a migration of this scale in smaller increments.