Engineering managers series part 5: business impact and knowing the why
In today’s post, we hear from Mathieu Bastian, Director of Data Science and Machine Learning, on the fourth component of our Engineering Manager framework: business impact.
Learn more about our Engineering Framework and it’s different components in Simone’s post on productivity, Rodrigo’s post on team health, Oliver’s post on stakeholder happiness, and Mattias’ post on systems health.
When I joined GetYourGuide in 2016, the team was composed of ~30 Engineers and a few Team Leads. At the time, we didn’t have Engineering Managers. The idea for the Engineering Manager role emerged when I was tasked with evolving our Engineering organization into a platform that could support 100+ Engineers. According to Reid Hoffman's blitzscaling terms, this meant moving from the Tribe to the Village, so I anticipated the task would be difficult. Nevertheless, just one month later we had our first Engineering Manager job posting live and a few compelling applicants!
Once we started interviewing for this role, I began to notice something odd. Many of the candidates came in with a strong executor mindset. There was considerable emphasis on the “what” of a project, but little mention of the “why”, and I felt myself bristle in response. At first, I couldn't really explain why I was reacting so strongly against this mindset. After all, many of the candidates had excellent track records shipping large projects and managing teams. But, every time we drilled into product decisions on their last project, the reasoning crumbled. After rejecting many candidates, I started questioning what I was looking for in my Engineering Managers. Was I expecting our Engineering Managers to be great Product Managers as well?
The answer came naturally when I took another look at our company core values and Engineering principles, which highlight ownership as a key expectation. As Udi mentioned, central to the Engineering Manager framework is the idea that the Engineering Manager's success is based on the success of their team(s). With that idea in mind, it makes sense to hold them accountable for things like business impact — an area traditionally owned solely by the Product Manager.
This doesn't mean the Product Manager no longer has a role within our cross-functional teams; delivering business impact is extraordinarily difficult and we still expect Product Managers to be in the driving seat. That being said, there are a lot of principles an Engineering Manager can follow on a daily basis to maximize the chances of their team's success when it comes to impacting the business. Let's look at the top 3 principles we've been applying at GetYourGuide:
Work on the right product
There's nothing more important for business impact than choosing which problem is worth solving.
In my role in building Data Products, I’ve realized this principle is even more critical because the data science development cycle is usually longer. A start in the wrong direction can be very costly down the line. This was illustrated to me recently as we developed a new search-ranking algorithm (one of the core ways our team drives impact) and the results weren’t where we expected them to be. During the retrospective we realized we hadn't consider the following questions:
How different will the ranking be? If we're just shifting around one or two positions, the ranking won't change enough to make an impact on the user experience.
How many people will be affected by this change? If all the top queries remain the same, it's unlikely the different ranking algorithm will produce real business impact.
This simple example shows that though a product improvement can look good on paper, it might not produce lasting business impact. In our case, team execution was good; the problem laid in the planning and prioritization.
Have a long-term vision and mission
Even if it takes two years to complete, defining a very ambitious long-term goal is worthwhile, i.e. "make mobile ticketing seamless."
Your mission statement should fit on a t-shirt — simplicity will help create momentum and support.
Choose a solvable problem
Make sure you're in control of your destiny. If your success depends on something you don't have control over or you have no idea how or when it will be solved, reconsider your position.
Choose a problem that will affect a large number of users
Are you optimizing on an edge-case? Research and make a quick calculation of the estimated impact before starting a project.
Most likely, building a product for your power users is NOT a winning strategy.
Align your goals with the company strategy and socialize with other teams
Build a small presentation (or press release) describing your strategy and present it to the relevant executives — this will make sure your work is aligned with the overall aim.
Socialize regularly with other leaders to avoid any duplication and foster synergies.
Start with the low-hanging fruit
Prioritize your work based on impact and ease and pick the projects that maximize both dimensions. Then, aggressively experiment to validate your hypothesis.
Use data and research — focus on the largest opportunities rather than the largest problems.
Set ambitious targets
Ambitious targets motivate the team, but it’s important to base those targets on realistic scenarios (e.g. insights from data, industry research).
Don't fire and forget — revisit your targets regularly based on recent learnings.
Measure the right metrics
A popular quote we repeat in the office is Peter Drucker’s, “If you can't measure it, you can't improve it.” Success needs to be well-defined with a clear way of being measured.
The key question is what we mean by the right metric? Can a metric be wrong? In short — yes. Since our goal is to drive business impact, a wrong metric is a metric that can't be tied to an overarching business goal. Let's take user engagement as an example. Intuitively, user engagement is a good thing to improve — it means more users come back to our website, right? However, if our goal is for those users to make a purchase we need to demonstrate that more user engagement equals more purchases. In an ideal world, we have used data to establish a causal relationship between our metric and the business KPI.
Start with the goal
A great metric is nothing more than a way to measure your progress towards a goal — start by making your goal(s) very clear.
A great metric moves only based on your actions
Choose a metric that you can ultimately tie to one of the core business KPI.
Include metric research in your planning and study the baseline beforehand. Be careful about the noise (e.g. seasonality, other teams, shifting customer mix).
Make metrics visible, and own them
Be transparent about where you stand and communicate often; there is nothing worse for motivation than denying the reality of the impact your team is having.
Create a simple (real-time) dashboard and put your metrics front and center in your ceremonies and in the office.
Together with working on the right product, being an expert at shipping is the most important skill to master to ultimately yield business impact.
Shipping often is like an insurance policy for subpar prioritization. Even if you're working on the wrong problem, the better and faster you can ship, the sooner you can course-correct. In addition, what you can’t learn through data analysis and research you can usually get through shipping.
At GetYourGuide, our core values reinforce a strong bias for action, so we often try to find the fastest way we can test a hypothesis and move on. Engineering Managers therefore need to juggle between building something very quickly to test a hypothesis and delivering solid and reliable systems.
Shipping something good now is better than something perfect later
Build the simplest thing that can possibly validate your hypothesis, but be aware of cutting corners that are detrimental to your customers.
Always Be Shipping - when you're not shipping, you aren't learning from your customers.
Hitting target metrics can be as simple as minimizing bugs or rollbacks — create a habit of using and testing your product.
Be the first-in-line to test new functionalities and give feedback.
Be paranoid about details
You are your team's safety net, but ultimately they should take care of the small details as well. Don't take on important topics like scalability or security solo.
Trust but verify — listen to your team and ask questions. You are the best-positioned to connect the dots.
By carefully choosing where the team deploys its energy, measuring progress, and mastering execution speed, Engineering Managers at GetYourGuide maximize their chances of delivering business impact.
However, it's also important to mention that chance plays a role. Indeed, only a fraction of A/B experiments yield positive business results, but we shouldn't be discouraged after an experiment turns negative as this often leads to beneficial learnings. Not learning anything from an experiment is the real failure! Building a team that thrives in an experimentation culture and has enough resiliency to move the hardest business metrics is definitely a challenge, yet a fun one. Checkout Rodrigo's post on Team Health to get some of his tools for building and maintaining a strong team in this kind of environment.