By Todd Robert November 10, 2014 Twitter LinkedIn Facebook
Todd Robert is Creative Director and also manages front-end development at Coalmarch, a full-service web agency that delivers creative, results-driven marketing.
We’re a full-service agency, so we have to deliver it all—a site that looks great, has good content, and ranks well. It’s the sum of all parts that make the experience work so well for our customers, so we care immensely about how the site operates once we hand it over.
The Nightmare on Thanksgiving
Black Friday is the biggest shopping day of the year, right before the newly-minted “Cyber Monday”, where revenue pours in from online sales. On of our biggest client was an ecommerce site, so we were proactive. Since we didn’t have past data to draw from, we boosted the site to prepare for an influx of activity. They even ran a couple commercials, driving a ton of online traffic to test the waters.
The site still crashed.
Our developers had to drop everything on day after Thanksgiving. It’s tough to be a manager and ask that of your team, and it’s even tougher to explain the lost revenue to your customer. It could’ve been avoided with better infrastructure.
Traditional hosting loaded up our developers with sysadmin tasks.
Before Pantheon, we had a development server here locally. We hosted with a popular cloud-based provider for whatever else we needed. But it was on us if we were running out of space, if there was a traffic spike, or if a website needed patches or updates.
As our agency grew and we started hosting more and more clients, it became pretty obvious that hosting was taking up too much of our developers’ time. We’re not a large company. We don’t have a full-time sysadmin. We wanted our team to be working on high-quality work for clients, but instead they were in the trenches of hardware maintenance and debugging.
We needed to build robust sites quickly and know they’d hold up during traffic spikes.
The process of setting up a website before Pantheon was pretty time-consuming. Server configurations, moving from dev to test to live environments, starting from scratch on the basic code—just getting Drupal up and running. We were burning time we didn’t have.
Because it was hard to predict traffic volumes for each client, we knew we needed to deliver them sites that could handle unexpected fluctuations. As the Thanksgiving scenario made clear, feeling prepared doesn’t always mean you are.
We chose a platform that plays well with Drupal.
So far, we’ve rolled out 52 sites on Pantheon, and we’ll do so for every new client from here on out. The platform is perfect for Drupal, which is the only CMS we support. Migrating Drupal sites to the platform was pretty painless, given the identical dev, test, and live environments and useful developers tools. Now we can develop the way we want and know that the infrastructure that powers our projects is solid.
Our clients get a much better value for the same price.
We handle all the inbound marketing and site improvements for most of our clients. They’d prefer that we maintain control over everything. Having our development, test, and live environments all together saves us a ton of time moving stuff around and making updates. Some clients have teams that prefer more access, and with Pantheon we can be flexible.
Even though their sites were more reliable and improvements got done faster, none of our clients had to incur higher monthly fees. They were seeing better performance and uptime, quicker responses from us, and greater ease in adding new features or sites. When you can offer that for the same cost, it’s an easy sell.
More billable hours + less money spent on hosting = success.
Most of the hours we used to spend on server maintenance now make us money. Our developers are so happy to be able to focus on what they love to do, knowing there’s a super-awesome architecture behind all of our sites. Clients get more productive hours of development for their money and we don’t lose revenue while we fix servers.
On top of that, our hosting account used to cost us $600 a month. Now that money can go back into the agency.
Happier clients, bigger prospects when sites run fast and never fail.
Some of our clients run large, complex websites. When their went from always a little sluggish to stellar performance, I was planning on sending stats from New Relic to support the good news. But we didn’t even have to. They noticed as soon as we migrated them and were thrilled.
We can go after larger clients with more complicated website needs now. We have absolute confidence in our infrastructure, and that’s something that comes across naturally when we’re talking to prospects and planning for big, exciting development projects.
