Introducing Analytics Product Manager
What is it and why do you need them in the organization
We all know that product development needs data and analytics. Be it for understanding the market/consumer needs, identifying pain points on funnels or user journey, to incorporate intelligence and predictive models for better consumer experience. Little do we realize that as much as product management needs data to scale and improve, data and analytics also need “product management” to ensure the deliverables are aligned with product needs and delivered in the most effective way possible.
Why do you need one?
You might wonder why we need a specialized data/analytics product manager. Why don’t the product managers directly assign the needs to the data analyst/data scientist? These product managers know what insights or models needed for the product, right?
The answer might be yes, but depending on how complex (several product streams/services, run independently) and mature (have solved hygiene problems and now looking on delight/efficiency problems) the product or company is, you might need a specialized analytics product manager. The main reason for this :
Complex data problems and emerging cross products solution opportunities
As the product grows, the data problems are no longer descriptive or diagnostic questions like “how was our conversion in the last X months” or “which user segments are dropping out of the funnel”. The data problems might be something more predictive and prescriptive like “what are the driving factors of customer satisfaction” or “which customers are going to churn next month”. The data solution needed is no longer a deck of charts and figures, but also a heuristics logic or data models to be integrated with the product itself.
Sometimes there are possibilities of building a solution for the whole platform, not only a single product. Requirements gathering with numerous stakeholders, and iterative analytics strategy to match with product release targets are needed here. With this complexity, it’s just natural to have someone to look at the problems holistically and propose a robust analytics solution, while maintaining communication with stakeholders involved.
What do they do?
Bridging the data (analyst, scientist, even engineers) team with product and business teams, ensuring the best data solution for product improvement.
They translate business problems into data/mathematical problems to be solved by the respective data team.
- Deliver analytics products. Just like any other product managers, analytics product managers are in charge of product deliverables. The only difference is the products delivered are analytics products, which varies from data instrumentation (tables, reports, metrics definition), insights generation (exploratory analysis — incorporating business analytics and statistical methods like time series, causal analysis, hypothesis testings), to data modeling (rule-based segmentation, machine learning models). They identify the requirements of the analytics products and work together with data analysts, data scientists, and data engineers in the team to ensure delivery. They can even define release iterations, just like agile development in the regular tech product development.
- End to end solution-ing. The analytics product managers are in charge of developing the roadmap of the analytics products, especially on data models which at the end of the day is used for product features or process automation. For this, they first need to identify baseline instrumentation requirements (data points, table structures, tools) and data retrieval methods. Then, they will be exploring potential analysis/data modeling approaches, which can be further discussed with data analysts and data scientists. After the approaches are defined, they work with data engineers to explore data pipeline solutions for product integration and usability. Once the grand idea is defined, the product manager will break down the tasks needed to be distributed to the team.
- Set up the governance process. The governance process is needed to (1) facilitate an effective working process for every stakeholder involved, especially for the internal data team and (2) ensure the standardized quality of products delivered. Works involved here include setting up a team structure (resources needed for different project types), cadences/syncs/checkpoints (with identified objectives and stakeholders needed, even external stakeholders — the product/business team), guidelines on executing work (i.e analysis framework or data model integration checklists).
The characteristics of Analytics Product Manager
Having Analytics Product Manager as a “bridge” between the product/engineering team and data team, there are some expectations of skillsets needed in this role.
- Technical knowledge. This is a must-have for an Analytics Product Manager. They need to understand basic statistics concepts and analytics/engineering frameworks in order to shape analytics products to be delivered to the product/business team. This includes SQL, data pipeline, visualization tools, statistics, and machine learning concepts. With this knowledge, they’re able to visualize the potential solution, resources needed, and delivery timelines.
- Understanding of process. As mentioned above, one driver of the needs of this role is to shape the cross-products analytics solution. They will need a helicopter view on product needs and data points/tools available. With that understanding, they can synthesize generic and reproducible/reusable products for multiple purposes.
- Facilitator/integrator mindset. There are a lot of organizations in which the data analysts/data scientists are very experienced and they are too focused on executing the work on their own. Analytics Product Managers are needed here to facilitate collaboration between them (conducting sprints, brainstorming sessions, problem-solving), and also with the product/business team. With the product team as “customers”, analytical products delivered should be aligned with the overall business goals and product roadmaps.
Final thoughts
When your organization starts to grow and you have more and more people in the data team, you might realize the redundant data initiatives happening and long delivery time for certain products. In that case, you might need Analytics Product Managers to organize the analytics deliveries, matching them with various product needs.
These people are the ones translating business problems into mathematical problems, envision the analytical and technology solution, and ensure the execution with the rest of the data team: data analysts, data scientists, and data engineers.
The needs of this role might not be realized yet as for most companies, this role is shared between product managers and data scientists/data analysts themselves. However, I see the emerging needs for this as data teams continue to grow inside tech companies, and asked to solve more and more complex data problems.
