Dear DevOps professional: Do you have a Hard Knock Life? I’ll admit, it’s pretty easy to get self-absorbed in our own 99 Problems and not appreciate how far we’ve come as an industry. For instance:
- Culture within Development and Operations organizations is shifting away from centralized command and control to embrace distributed empowerment of teams and individuals.
- Automation is being embraced as a key mechanism to maintain guardrails while scaling the organization and product investments.
- Lean principles are helping engineering teams focus on value from an outside-in (customer-centric) perspective, instead of a technology-centric worldview.
- Measurement is driving more informed decision-making by forcing more rigor in modeling what improvement should look like.
- Sharing information and experiences is easier than ever before, ensuring transparency and load-balancing instead of having tribal knowledge reside with a few people.
Even if you’re not quite there yet within your IT Operations / Engineering bubble, I implore you: Brush that Dirt Off Your Shoulder and realize those DevOps principles (the CALMS of DevOps) that you’ve worked so hard to implement can help in other parts of your company, too. Let’s look at your Marketing organization as an example.
Give It 2 Me
Marketing tech (or MarTech) has come a long way to help marketers improve engagement with customers. Traditionally, especially for core systems like a CRM, IT would be called in to build, deploy, and operate a solution custom-tailored to particular (and ever-changing) business needs. But let’s be honest, marketers: IT doesn’t understand your domain at all. You want to change the keywords in a particular ad campaign? Or ensure prospects aren’t being bombarded with too many drip emails? File a ticket. Oh, it’s time-sensitive? File a P1 ticket—the one with the two-week SLA.
Today, digital businesses promote agility by empowering teams on the front lines (aka the marketing team) to make informed decisions quickly and giving them simple, customizable tools to put them into action. For SaaS companies, the first step is often the marketing website, which frequently runs on shared infrastructure and customized by a shared development team. Now, there is a trend toward forming Marketing Ops teams, who have full-stack ownership and align to the goals of the broader marketing organization by leveraging the organization’s domain expertise and applying technology appropriately.
If you’re in an IT Operations org supporting Marketing today, work on simplifying and training your counterparts within Marketing to actively build that culture of empowerment, and ultimately Give It 2 Them (by that, I mean giving them ownership of the website entirely).
Can’t Knock the Hustle
As a software-developer-turned-product-manager, it’s pretty interesting to see what lengths people will go to in order to solve their problems: They copy, paste, query, refresh, communicate, and schedule—all manual activities that take up quite a bit of time. In applying a technology mindset to the Marketing domain, the most obvious way to drive efficiency is by introducing some semblance of automation.
Automation lowers the barrier to entry for a particular process and makes it more predictable and less error-prone. Even simple changes, like setting up threshold alerts for social media activity instead of constantly refreshing Twitter or LinkedIn, can save a ton of time. This type of model change (push vs. pull) is a solved problem in IT Operations: Every infrastructure or application monitoring tool has had that capability for decades. Applying the same principles to digital signals might be newer to MarTech, but the approach can certainly be re-used with success.
Everyone can benefit from automation, not just those in an ITOps discipline. Just think: Wouldn’t it be great if you could send social media activity alerts as digital signals to PagerDuty so your team can coordinate real-time responses and correlate signals across tools? And if the threshold was a bit too sensitive, you could discuss that as a team in a postmortem and maybe record a note using a runbook in case it triggers again for the next person on-call.
In Lean parlance, this type of automation helps to “eliminate waste” from a process. Another key principle of Lean centers on delivering value to the customer by specifically mapping out the parts of the process to that value stream. For example, if you discover that sending an email out to sales reps about the leads collected at a conference doesn’t provide value (since the reps look in Salesforce anyway), don’t automate the email! Remove that step from the process entirely. Hustle is always appreciated, but it’s only useful when it results in value.
‘Empir-ical’ State of Mind
One of the nice side benefits of introducing automation is that it forces you to sit down and understand the overall process, no matter what that process is. But let’s step back a bit and ask: What does success really look like? If the process is faster, how much faster is it? Do you have a goal in mind?
I’ve loved watching ITOps trend toward being more intentional about what you’re measuring and how that reflects in alert design (I mean, there’s an entire conference dedicated to it: Monitorama). Measurement of the various metrics within the Marketing domain—including the number of qualified leads, click rates, ad campaign spend, etc.—is really only the first step, but it’s a critical one to understanding whether something you’re trying to improve is actually improving.
Another important aspect of measurement is that it forces you to prioritize which metrics really matter. For years, ITOps professionals have focused on system metrics like CPU, disk utilization, network throughput, cache size, etc. Those metrics can be helpful, but they don’t really represent the customer experience or the health of the business. Is there an equivalent upleveling that needs to happen with Marketing metrics? Perhaps. But the Marketing domain benefits from being far closer to the business than abstract applications or infrastructure. Regardless, being data-driven is a core tenet of DevOps that applies just as well to Marketing—basically, if it moves, graph it.
H to the Izz-O
Holding on to information that might be beneficial to a colleague is not typically done overtly. Sharing is a core principle of DevOps because it’s an intentional act done for the benefit of the team. Chat tools like Slack, Stride, and Microsoft Teams have made it simple to have a conversation in the open and connect people across geographies and timezones. The key to using them optimally isn’t in point-to-point (person-to-person) communication. It’s easy to forget that chat rooms have been a familiar tool in the toolbox for ITOps professionals for a long time, hearkening back to the days of Internet Relay Chat (IRC). For marketers (and other business disciplines), moving past the private 1:1 conversations is one cultural step toward better sharing.
Lest I lead you to believe that sharing is only tied to chat, there are many other ways that a “sharing culture” must emerge. To help your Marketing team become more effective and efficient, you can leverage the tools you already have (e.g., a company wiki, a Slack channel, PagerDuty, etc.) to become information radiators. You can also:
- Establish common processes for activities such as onboarding or new-hire training.
- Circulate best practices using methods like lunch-and-learns.
- Spread key learnings from a customer call via a retrospective / postmortem.
- Ensure common ground in conversations by providing links that everyone can access.
As you can see, you don’t have to be a Renegade with DevOps to uncover ample opportunities for bringing the same helpful principles outside the bounds of your comfortable IT Operations / Engineering organization.
However, before you go ply your newfound powers of DevOps good within your Marketing group, be aware that marketers have been coming at this from the opposite direction, too. Stay tuned for the inside scoop from that perspective coming soon! (Though, I’d bet there will be far fewer Jay-Z references—in case you were keeping count at home, there are at least eight great songs in here that you should go listen to.)