Rigor’s “The Best in Performance” interview series offers the chance to listen in on conversations with web performance industry experts, thought leaders, and technologists as they discuss current trends, challenges, and lessons that impact the performance and APM space today.
These insightful videos will provide you with information from industry insiders that you can use to plan for upcoming events, start conversations, and make an impact in your own organization.
Check out highlights from our conversation with Rich Howard, then view full episodes in the series (including interviews with Cliff Crocker, Buddy Brewer, and Alla Gringaus). Be sure to subscribe so you’ll be alerted when new episodes go live.
Rich Howard is the founder and CEO of Optimal, a company dedicated to improving the performance of websites and mobile apps. Optimal is located in San Francisco, and the company provides services to clients in North America, EMEA, and Asia Pacific. Prior to founding Optimal, Howard worked for companies such as Sephora and Williams-Sonoma, and he served as the lead architect on Vodafone’s global optimization team.
Howard sat down with Rigor’s CTO Billy Hoffman to discuss best practices for holiday and peak-load readiness from a user experience and web performance perspective. He goes into detail covering the steps that need to be taken to ensure that a website or application is prepared for an influx of traffic, from scoping and testing to setting expectations and creating a contingency plan.
Rich Howard on…
- Peak-load readiness.
- Practical planning.
- Stakeholder involvement.
- Performance testing.
- Contingency planning.
- Proven techniques.
- Lessons learned.
Q: What in your background helped you gain a stronger understanding of peak-load readiness?
When working on a project for the Man Booker Prize in 2007 (now known as the Booker Prize), Howard saw first-hand how a heavy-load day could impact a website. The prize has three key announcement dates on its yearly calendar – the long list, the shortlist, and the winner. As Howard worked on updating the website, he and his team discovered that the site could not handle the amount of traffic that flowed in on the first of the three key dates. This resulted in users complaining and the site itself grinding to a halt.
“The challenge for us at the time,” he says, “was that we had no previous traffic data, so we didn’t know what to expect when we were facing these loads.” He adds that performance “wasn’t baked into the original design of the website,” making it difficult for the site to handle the heavy traffic load.
To solve the issue, Howard and his team implemented several technical updates, disabled non-essential parts of the site during the peak event, and introduced additional testing to a readiness plan. “All of this combined really saved the day,” he says.
Q: How can a business start thinking practically about readiness for peak-load events?
“It doesn’t matter if it’s an ecommerce website or a site within a different realm,” Howard says. In all cases, it’s critical for conversations to happen consistently between business teams and technology teams.
To start the process of thinking about holiday readiness, he recommends that questions be asked such as, “How many people are we sending out to? How many do we expect will be engaging with this campaign? What do the numbers look like?” Then, use that data to determine whether the site can handle that volume of traffic and to start planning for how to prepare the site.
Q: How can stakeholders work together to build a readiness plan?
Howard recommends that business and technology stakeholders align to understand what they’re preparing to face during a particular event. These teams can work together to uncover existing technical limitations, identify current plans to make site upgrades or optimizations overall, and ensure that all parties involved know what will be expected of them as part of the holiday readiness plan.
Howard explains, “Sometimes, the business and technology stakeholders may be speaking a different language, but they’re all trying to achieve the same goal – to have a successful event where performance is maintained and the users have a great experience.”
[bctt tweet=”Business & tech stakeholders are trying to achieve the same goal-to have a successful event where performance is maintained & users have a great experience. – Rich Howard #webperf #thebestinperformance” username=”TeamRigor”]
Q: Once you have everyone on the same page in terms of expectations, how do you determine the scope of the readiness plan?
At this stage, it’s important to understand the traffic patterns of the site as well as what the expected load, volume, and scale are going to be during a peak-load event. “I’ll often deep-dive into a customer’s analytics system to figure out what particular funnels or journeys are being used to understand which types of pages are going to be heavily trafficked,” Howard says. From there, he will calculate a load model.
Howard also will run a system audit to get a view of which measurement and monitoring systems are already in place, to identify places where there are gaps, and to make recommendations for improvement. Then, he will look at how all tools are configured and deployed in each environment to determine if they are providing the necessary data.
Q: How do you suggest people approach performance testing as they prepare for a holiday or peak event?
Seasonal readiness testing involves both the front and back end. On the front end, Howard explains, “there’s a lot of testing we can do to understand how the pages are built and what we need to do in order to optimize them.” For the back end, Howard notes that load and scalability testing should be incorporated and then combined with APM data to give an understanding of what is happening within the system.
He notes, “I like to create a combined strategy whereby we’re looking at data from different tools, then coming up with candidates for optimization based on the data coming out of those tools.”
Q: What are your recommendations for creating a peak-event contingency plan?
Contingency plans have a “people aspect,” a “process aspect,” and a “technology aspect,” says Howard.
[bctt tweet=”Contingency plans for peak-load events have a ‘people aspect,’ a ‘process aspect,’ and a ‘technology aspect.’ #webperf #perfmatters #thebestinperformance” username=”TeamRigor”]
From a people perspective, Howard explains that teams need to work out a map of all stakeholders who will be involved during a peak event, including who will receive alerts and who will be part of an escalation path.
From a process perspective, Howard says it’s important to review what happens after an alert is triggered. He says that teams should ask “How are we going to get together to resolve this issue that we may be facing? How do we triage it, how do we assign it to a team, and how do we react quickly to a possible issue?”
From a technology perspective, Howard says that teams should be looking into how to build systems that are more resilient overall. Teams should ask, “If a system fails or a component within the system fails, is there a way we can recover from that and still maintain the customer experience, even though one of the pieces is currently not functioning?”
Q: What are some techniques you’ve seen businesses put in place during peak-load events?
Howard says that he’s seen businesses choose to:
- Disable non-essential functionality, such as on-site search tools
- Scale back or remove third-party systems that may not be crucial to end-user flows
- Temporarily remove A/B multivariate testing frameworks from the site
- Make technical changes like increasing the length of cache headers within their particular assets
Q: How do you approach retrospectives for peak-load events?
“Chances are you’re going to have a similar event in the future,” says Howard, “and there are always things that you can learn from a previous event.” He recommends holding a retrospective as soon as possible after a peak-load event. “It’s very easy to lose the flow. Everybody goes back to their normal everyday roles after the big event, and so it’s easy to forget what exactly happened,” he notes.
When holding a retrospective, Howard says that it’s important to involve members of all teams across the business, rather than just technical teams. Every team will bring a different perspective to the table. “It’s important that they each have a voice so they can share the things that could be improved upon in the next event,” he says. “We definitely need to have the right people in the room to have those conversations in the spirit of saying how we can work together better in the future to have even more success with our events.”
In connection with the retrospective, Howard recommends creating a collaborative document or wiki page where members of various teams can add their data and information for future reference. This can be sectioned by discipline, and there can also be a section for more general issues that affect multiple teams.
He adds that it’s important to review the data retention policy of any key systems being used for monitoring and measurement. If a system only retains data for a short period of time, it’s critical that someone is tasked with grabbing that data and archiving it if it needs to be reviewed again at a later date so that it’s not lost.
Q: Is there anything you’ve learned now about testing that you wish you’d known earlier in your career?
The biggest lesson Howard says he has learned is, “Traditional performance testing really doesn’t look at the entire system as a whole.” He notes that it is important to test and complete all user journeys, when possible, and to review and manage all third-party integrations. “Being able to performance test with [third parties] in the mix can be difficult and challenging,” says Howard.
[bctt tweet=”When it comes to performance testing, you need to understand the system as a whole, including all integrations and user journeys, to get a complete view. – Rich Howard, CEO of Optimal #webperf #perfmatters #thebestinperformance” username=”TeamRigor”]
To prepare for peak-load events including, but not limited to, Black Friday and Cyber Monday, Rich Howard recommends that businesses take a multi-step approach. This includes scoping and testing, identifying candidates for optimization, creating a contingency plan, and then running a retrospective after the event. By reviewing all data and learning from both successes and mishaps, the business and technical teams can continue to refine and manage their holiday readiness plans for future success.
Check out Rich Howard’s full episode of The Best In Performance video interview series – as well as episodes from other industry experts. And don’t forget to subscribe so you can be notified about new episodes as they’re released.
The Rigor platform can be used to understand critical user journeys as businesses develop their holiday readiness plan. To learn more about how Rigor can help your business prepare for peak-load events, request your free trial.
How has your business implemented holiday readiness plans for your website or application?