A Culture of Partnership

Two sets of partners in a three-legged race all smiling. One girl has pigtails flapping around in the wind.

A Culture of Partnership

During my time leading an analytics and data science team, I spent a lot of time thinking about how an ideal analytics team should operate – how the team should work together, how the team should prioritize their work, and how the team can most effectively partner with the broader organization to generate business value.

I believe that for an analytics team to be effective, the team must develop a strong culture of partnership in order to actually drive business value. An analytics team that operates in a silo, despite being filled with super talented analysts, engineers, and data scientists will often end up spinning its wheels and failing to impact the bottom-line of the organization.

Why partnership?

In most companies, analytics and data teams are cost-centers rather than revenue drivers for a business. In order to generate as much value as possible (and avoid getting your budget slashed) it’s critical that your analytics team form deep and long-lasting partnerships with others throughout the organization. The benefits of these partnerships are myriad, but to start they can:

  • Give you and your team additional insight into underlying business complexities
  • Provide you a source of new data that you can feed into the data quality flywheel
  • Stimulate questions for interesting research or product opportunities
  • Evangelize for the value that your team provides (since most of your work doesn’t directly impact the bottom line)

While it may be obvious that it’s important to have a culture of partnership, it’s devilishly un-obvious how to go about building a culture of partnership either when building a team from scratch or when trying to change the culture of an existing org. In this post I’ll talk about how I think about building a strong culture of partnership that can help an analytics team identify the most important opportunities in the business and work effectively with key stakeholders and partner teams to drive value.

I’ll break this post into 3 parts:

  • Team values
  • Field research
  • Building relationships

Team values

When I started building the analytics team at Harry’s, one of the first things I did was I sat down and wrote out what I wanted the core values of my team to be – the first two things on the list were that an analytics team should be humble and helpful.

These two values are critical to building an analytics team that exhibits a culture of partnership (and, in turn, building a high-functioning analytics team). Humility is how analytics practitioners should engage both with potential partners but also with their own work – hubris is the enemy of good research. Helpfulness is the best tool for helping the analytics team build effective partnerships with other teams that might not always understand what the analytics team actually does.


Analytics teams work best when they

  1. have the support of the rest of the organization behind them and
  2. work to truly understand the use-cases, problems, and expertises in the rest of the business

One of the joys of working in analytics is that you get to work on a variety of different problem domains – moving from marketing, to product development, to software engineering. However, that often means that we are dropping into a new problem space without complete knowledge of the domain or of the history of the team we’re working with – humility helps the analytics team approach new problems knowing that they are not the experts.

When an analytics team is humble, it makes the partners feel like the experts in their domain (which they are!) and it can help stem some of the turf-battles that can occur when an analytics team drops into a problem area with guns blazing ready to shake things up.

In addition to putting their partner teams at ease, being humble helps the analytics team do better work – the work of analytics teams often involves a mixture of complex engineering and statistical work with a dash of assumptions thrown on top. Always keeping in mind the ways that your analysis could be incorrect, or the ways that your software could generate wrong results is critical to reducing the frequency and minimizing the impact of errors.

Of course, being humble does not mean refraining from ever giving advice or making suggestions. There will surely be times when it’s appropriate for the analytics team to jump in and use their expertise to drive change. The value of humility is that it 1) makes the partner team more receptive to the feedback or suggestions when it comes and 2) increases the chances that the suggestions are well-informed and actionable for the partner team.


Analytics teams should strive to help their partner teams (and be sure to share the glory in whatever they partner on). In my experience, a little dose of helpfulness goes a long way – even when the helpfulness doesn’t directly relate to the project at hand. A good way that the analytics team can often help out partner teams is with simple workflow automation – many times I’ve taken a slow and cumbersome excel workflow off of someone’s hands and sped up the process dramatically by re-organizing the excel file and moving some of the heavy computation into SQL. This made my partner’s life easier (saving them 15 painful minutes every morning) and made them especially keen to help me when I need their assistance with some other project.

When an analytics team is both helpful and empathetic, other teams will be lining up to work with the analytics team. Even when you need to ask a team for something (maybe you need them to upload some data into the warehouse regularly) they’ll know that you 1) value their time and so wouldn’t be asking for something if you didn’t need it and 2) that you’ll be around to do them a favor in the future when they need it.

It’s critical that the analytics team be seen as both helpful and mindful of the burden you’re placing on other teams. If you come with a request or an ask from another team, it’s important that you let them know you’re interested in partnering with them to make the process as easy on them as possible to minimize disruption to their workflow.

Field Research

One of the best ways to encourage your analytics team to both empathize and partner with other teams is to have them actually physically spend time with potential partners. Have them attend their weekly meeting (and keep their damn mouth shut), have them do “pairing sessions” where they watch them do their work and learn about both their internal thought process as well as the tools they have and the limitations of the systems they work in. They probably have stakeholders, incentives, and tooling limitations that your analytics team might never have thought of before.

After a session of field research, it’s much easier for the analytics team-member to

  1. respect the challenges of this person’s job
  2. appreciate the subtle complexities of the constraints that might make potential partnerships or off-the-cuff solutions more challenging than one might originally think

Additionally, often these periods of field research can lead to exciting new product ideas that can lead to both improved workflows or novel data science products. For example, while working with your finance team you might come up with ideas for products that can help them with forecasting, or when working with the marketing team you might come up with new ideas about how to more efficiently test combinations of landing pages and advertising copy.

Building Relationships

This technique is both one of the simplest to understand and the hardest to enact. People on the analytics team should invest in building personal relationships with their partners in the rest of the organization. These personal relationships will go a long way toward making the analytics team more effective by greasing the wheels when your team needs to work with one of these partners. Building personal relationships in the workplace can be tricky, and different people will approach these relationships differently. To be clear, I’m not saying that you need to encourage your team to go out to booze-filled happy hours. Rather, the types of personal relationships that matter are the kinds that can be built person-to-person over a shared lunch in the office, or during a walk for a coffee (or tea!).

Building relationships with people can be as simple as asking about their lives and interests outside of work – how are their kids? were they able to fix that leak in their apartment? how’s the commute from Brooklyn? These sorts of conversations actually go a long way toward building robust relationships that make working together both smoother and more fun. Building up that level of comfort so that people on diverse teams across the business feel comfortable coming to your team with questions is incredibly valuable. Even more valuable is having the level of comfort so that you and your team can tell someone “no” because their ask isn’t prioritized without suffering any dramatic blowback – that sort of trust and good-faith is most easily built through personal relationships and mutual respect. A worthwhile investment for sure.


Building a culture of partnership on your analytics team is crucial to maximizing the impact your team can have. Without it, an analytics team can end up siloed, shouting the results of their research into the void.

If you’ve had success (or troubles!) building a culture of partnership, I’d love to hear about them. Shoot me an email or come join our slack channel!

Michael is a data strategist and entrepreneur working on innovative data products. He has experience applying econometric research methods to fields including environmental economics, child welfare policy, and medical treatment efficacy. Michael is the former director of analytics at Harry’s where he worked to empower the organization to “make better decisions faster.” In his spare time, Michael drinks too much coffee, reads, and pets dogs around Mexico City.

Site Footer