Glue Work

A woman holding a glue gun with lots of charts and graphs in the background stuck to the wall. The glue is everywhere.

If you’re lucky, you know what it will take to get your next promotion. Maybe you need to work on communicating your analyses at the C-level instead of primarily to functional managers. Maybe you need to finish a complex project like a predictive churn model. But there is lot of work in every analytics role that just has to be done, even if it doesn’t seem to help you get from where you are to where you want to be. You have ad-hoc requests coming in, you want to test out a new tool that looks like it will really improve productivity, and you spend an awful lot of time in cross-functional meetings just trying to stay plugged into what is going on around the company.

Tanya Reilly created the term glue work to capture “the less glamorous – and often less-promotable – work that needs to happen to make a team successful.” Glue work is valuable – without it, projects fall apart – tasks get dropped, teams miscommunicate, and it is just harder to get things done. However, there is a difference between valuable and valued. If there’s nothing in your job description or career ladder that addresses this work, when it comes to promo time, it can be hard to get credit for the time you spend on these tasks. If glue work makes the team function more smoothly and effectively, but it is not valued, then that is a problem.

Ultimately, team management should be responsible for aligning valuable and valued in their team’s work. Unfortunately, there’s often a disconnect. If you’re a manager, there’s a lot you can do to help that situation. But even if you aren’t (yet), you can take tangible steps to set yourself up for success (and hopefully a promotion). Let’s dig into it.

What Does Glue Work Look Like in Analytics?

Analytics has become much more technical in recent years, and Reilly’s descriptions of glue work resonate in some ways. But analytics is by its nature more collaborative and cross-functional than engineering, and I think that yields some meaningful differences.

Engineering is a well-defined discipline, with established norms and practices – whereas in analytics we’re in the midst of a period of transition where the dominant tools, processes, and expectations are shifting. We’re adopting some of the best practices of engineering, but it’s a work in progress. Most of us don’t have (and maybe don’t want) a product manager to sort through competing stakeholder requests and determine priorities – we work closely with our stakeholders and do that ourselves. Our teams often include folks with a pretty wide range of technical skills. There is a lot of variation in the work we do, and fewer clear rules on progression.

Engineers also generally report into other engineers all the way up the org chart, so there is reasonable alignment on what success looks like for an individual contributor. On the other hand, an analyst may report into a manager who understands the nuances of their technical and non-technical contributions, but the next level manager could be an engineer, a finance leader, a marketer, or anywhere in between. As a result, there isn’t always good alignment on what the most important parts of Analytics work are. Your manager almost certainly has to defend your promotion at that level, and may be actively competing for promotion slots – so clarity and visibility of what is promotable becomes even more important.

From the small sample of teams I’ve been on, there may be as much variance between analytics teams as there is between analytics and engineering. But defining what glue looks like on your team is the first step to combatting it.

A Few Examples

The short answer is, what is promotable varies depending on what is (and isn’t) in your job description or on your career ladder. Glue work might include any of the following:

  • Cross-functional collaboration: building a culture of partnershipcommunicating well, and training stakeholders on tools or data sources
  • Ensuring smooth analytics team functioning: interviewing and then onboarding new team members, unblocking your teammates, designing new templates or processes, planning offsites
  • The necessary but less exciting parts of our jobs: responding to ad-hoc questions and triaging inbound requests, documentation, scheduling meetings

Sometimes within analytics teams, it’s work that crosses job descriptions. It can be:

  • The analyst who implements automated testing
  • The engineer who reaches out directly to a sales rep to understand how the data they scrape ultimately gets used so they can improve the quality
  • The junior team member who remembers what a debacle their onboarding was, and creates a new onboarding plan for the next hire

And sometimes, technical data folks report into or have to pitch their promotion recommendations to a less technical leader. In that case, highly technical work can actually become glue work, including:

  • Designing processes and systems that scale well
  • Implementing monitoring and alerts for data freshness or quality
  • Improving query performance
  • Anything that is not immediately visible to end users of a reporting or visualization tool

Why It Matters

All the work listed above is important and valuable. And all of it belongs in an analytics team. It really shouldn’t be glue work – it is the work. The alternative is wasting time on analyses that aren’t aligned with business priorities, team members getting stuck, accumulating technical debt, losing trust from stakeholders, and maintaining painful manual processes. I lead analytics teams, and I want my team to be rewarded for glue work so we can all be more successful – and spend more time on work that will make the company more successful.

But it also matters a lot because the distribution of glue work is not equitable. Women are more likely to volunteer for non-promotable tasks at work (much like they do at home, but that is another story). Representation of women in analytics varies across specialties – women and men are close to equally represented among analysts, but less than a quarter of data scientists and engineers are women. Only 24% of analytics leaders are women. Tackling glue work is a real, tangible way we can help women advance and to work toward more equitable gender representation in analytics leadership.

If you don’t have any women on your analytics team, well, think about that the next time you’re hiring. But in male-only groups, the same men tend to volunteer time after time for non-promotable tasks. So regardless of the gender makeup of your team, a small number of people on your team are probably doing this work without being rewarded for it.

The Meta-Story

Arguably, analytics itself is glue within many companies. In most companies, the analytics team is not responsible for a revenue target. The team may not own any OKRs. The impact of analytics is subtler. Good analysts are multipliers, enabling the company to identify and focus on areas where changes can have the most impact. We don’t build features, we don’t close deals, but we shape the processes, tools, and cultures that enable companies to be smarter and faster across the board.

Yet, there is often confusion about what analysts are – they just write reports, right? Data science has the opposite problem – folks on the outside think it’s really exciting, but the reality is that it’s exciting occasionally, and boring a lot. In either role, when you tell someone that the “simple” request they asked for is really multiple days or weeks of work because the data isn’t yet available, isn’t well-defined, or needs significant transformation work, their eyes may glaze over.

These disconnects represent a misunderstanding between the company as a whole and the analytics team about what is valuable, and what is valued. Clearly identifying and communicating why different types of work is valued within the analytics team sets the stage to more clearly communicate the value of the analytics team itself to the rest of the company.

So What Do We Do About It?

The problem of glue work is fixable. If your team is doing work that helps improve the likelihood of success on projects, that work needs to be explicitly valued.

Managers

If you are in a position to drive systematic change in your organization, you can help align valued and valuable:

  • Figure out what your team is doing that is not written in their job descriptions. Watch them, and ask them – about themselves and their teammates. Separate the list into work that requires an analyst’s skills and improves the output of the analytics team, and work that doesn’t.
  • If the work is truly valuable to your team’s mandate, write clear career ladders that reward what needs to be done – including glue work. This might include documentation, relationship-building, onboarding new team members, and building subject matter expertise. I have a lot more to say on this – enough to fill another post, coming soon.
  • If the work just needs to be done, but is not particularly relevant to your team’s mandate, then find a fair way to assign it. There will always be tasks that are non-promotable but need to get done, like scheduling meetings or planning team off-sites. Either you do that work as team manager, or explicitly assign a fair rotation of these tasks within your team. Don’t wait for volunteers.
  • Understand where your team members want to go. Are they trying to get from Analyst to Senior Analyst? Do they want to get from Analyst to Engineer? Or do they ultimately want to move from Analytics into Product or Marketing? Not everyone wants to climb the same ladder, and sometimes work that is not explicitly promotable in analytics is still helpful in getting that person where they want to be. Just make sure those choices are clear to your team members as you work with them on development plans.
  • If there is enough of it, what used to be glue work can become its own role. That can be a rotating role, like a daily triager who answers questions, acknowledges inbound requests, and begins the process of estimating and prioritizing. In can be a whole new analytics role, like an analytics engineer who fills the gaps between analysts and data engineers or a data product manager who excels at coordination, prioritization, and stakeholder communication. For people in those roles, excelling at these tasks is their core focus and helps them achieve their career goals instead of being a distraction from what they “should” be doing.
  • Manage upward. Just because you value what your team is doing, doesn’t mean you’ll be able to push through their promos when review season rolls around. Make sure that your manager, and anyone else who needs to support a promotion, understands the less visible parts of what your team does. Make time in 1:1s with your manager to brag about the impact of your team’s work. She may not understand what it means to refactor a Looker model to improve performance, but she’ll understand what it means to cut the load time for her most-used dashboard in half. Make sure she knows exactly who did that work. This helps your team, but also helps you with the meta-problem of analytics itself as glue by demonstrating the less visible ways your team impacts the company.

Non-Managers

Maybe you are not in a position to change the system – yet. So what can you do to make it more likely that you’ll get promoted (and eventually be able to make those changes yourself)?

  • Understand where you want to go. If you are aiming for a promotion within analytics, talk to your manager to understand what it takes to get there. If you have another goal, find a mentor who can help you create a path. “Promotable” is only meaningful when you know what promotion you’re aiming for.
  • Figure out what you are doing that is not on that path. Talk to your manager and/or mentor about those tasks – find out if that glue work will help you reach your goals, or if she truly doesn’t value it. In a perfect world, your manager sees the value in your glue work and adjust the career ladder to reflect it. But the world isn’t perfect, so …
  • Tell the story of your glue work. If your manager doesn’t initially consider your work progress toward a promotion, help her really understand it. Explain how long the work takes, and how it adds value. Quantify the impact if you can – have ad-hoc requests declined since you started spending more time training people on internal tools? Has the new hire become productive 3 weeks faster because of your new onboarding program? Make it real to her.
  • And if she still doesn’t value the work? Stop volunteering for it. That sounds extreme, and it can feel like you are dropping the ball. But if it is important, someone else will step up to share the load, or your manager will figure out that it did matter, and assign it (hopefully fairly).
  • What if someone asks you specifically to do glue work? Tell them you have different priorities, and suggest someone else who can help share the load.
  • Bonus points: help others who over-index in glue work realize what they are doing, and ask whether it serves their goals. Sometimes small things highlight inequity better than big, substantive conversations. Are women the ones who always cut and serve the cake at birthday celebrations? Point that out to your women colleagues, and together, you can all stop doing it. Your male colleagues probably don’t even realize it’s happening, but I promise – they know how to cut a cake too.

Glue work puts the health of the team ahead of individual accomplishments. It is core to the success of individual projects, but also to maximizing the impact of the analytics team overall. Explicitly valuing glue work and not relying on volunteers to do it helps team remain sucessful while leveling the playing field and improving gender equity.

I would love to hear what glue work looks like on your analytics team, and what you’ve done to address it. Send me an email or come have a chat in our slack channel!

Caitlin is Head of Analytics at Trove Recommerce, where she helps brands create circular retail models at scale – buying back and reselling their products to extend their useful life. She previously led data teams in crowdfunding and self-publishing. When she’s not optimizing single-item pricing or operations, she’s usually drinking a cup of tea and watching her chickens peck around the back yard.

Site Footer