As an early data hire at a fast-growing company, one of the first things that you’ll likely encounter is a backlog of questions from employees. This backlog, alongside all the other tasks associated with reporting, maintaining data and creating new pipelines can feel extremely overwhelming. These long lists of tasks and questions prevent many data teams from being proactive about analysis. Without a proper process in place that allows analysts to enable self-service, managing data requests proactively can feel like a never-ending battle against a current. Additionally, data teams that constantly need to answer the same question are not creating processes that help them manage the scale and complexity that companies experience as they grow. By the end of this post, we should’ve shed some light on the following questions:

How to start a data requests process as a new data hire?
What's a data request?
How should I work with stakeholders?
What does a early data person do at a startup?

Below are some of our tips on how to manage the data requests backlog. We believe, that great data teams should be proactive. They should adopt tools and processes that ensure that the data team never has to answer the same question twice. We hope this step by step list is helpful to all data teams that want to improve their efficiency and reduce their workload in the future:

1. Set expectations about the data requests workflow with reasonable timelines

The first step of setting expectations around a team is communicating the way that your team is working to the stakeholders in the company. We suggest adopting an Agile workflow with weekly or bi-weekly sprints. Although your Scrum team may be small at first, setting expectations about when certain requests will be answered with the sprint methodology can be helpful. With scrum, a product is built in a series of iterations called sprints that break down big, complex projects into bite-sized pieces. 

Teammates can explain why and how quickly they need certain items to be completed and the Scrum team can tackle new projects every week. If something comes up during the week that has higher urgency, the team can switch tasks but needs to re-evaluate timelines for unfinished projects. 

The most important part of this step is making sure that people are aligned with the way that you work. Too many times, data teams are bombarded by incoming requests and unrealistic expectations about when something will be completed and the complexity of a task. This won’t solve the problem of unrealistic timelines completely, it should help some employees think proactively about how they work with the data team. 

2. Define your requests workflow

Most teams we speak with choose to manage their data requests directly in Slack or Jira. While this is a good place to manage their requests, it usually doesn’t scale very well and causes teams to answer the same question multiple times. Data teams should aim to make old answers to common questions searchable and accessible so that they aren’t answering the same question multiple times. Additionally, if a certain dashboard or metric is difficult to find, they should think about ways to make it more legible for their customers. Data teams should look for patterns in their data requests and try to help employees by solving the problem at the root.

By managing a transparent and inclusive data requests process, data teams can achieve a similar outcome to support teams. They can start to data requests with good documentation and common FAQ’s and they can start sending more time servicing the difficult question or the tasks in their work that require more effort. Data teams that think ahead and want to spend less time getting pinged on slack in the middle of the day should think about adopting a data knowledge management tool like Secoda to solve their problems. By defining your data requests workflow, you can ensure that teammates are adopting the best practices when asking your team questions around data. Next, teams should make sure employees use a request template when they ask questions to prioritize work.

3. Create a requests template

We’ve created a data requests template for teams that are looking for a better way to manage inbound questions. Below is the template. Feel free to copy the template and use it in your team's workflow. The template and data requests can be automatically created with Secoda.

  • What is the business question you are trying to answer?
  • What is the impact of this question and how will it help the company?
  • Who will be using this data?
  • What time frames are crucial here? (Example: Monthly, weekly, daily)
  • What is the visualization you are trying to create?
  • What interactions/drill-downs are required? (Ie. the type of use, revenue amount etc.)
  • Are there any other details we should know about this data request?

Having a template for all questions allows employees to test their assumptions before submitting a question. Combined with data request context, a business is less likely to provide bad or incomplete data, which then reduces the risk of making bad business decisions.

4. Automate repetitive data requests

Data teams that take the next step with their data requests process can start to think proactively about data requests. Customer support teams have been deflecting inbound questions for years using tools like Intercoms knowledge base and Ada automated customer support chatbots. Smart data teams realize that they can do the same. 

Data teams can automate and deflect common questions with tools that allow them to document data requests in the same place teammates are looking for answers. Secoda lets data teams integrate their knowledge repository directly into Slack and lets employees ask questions using the pre-made templates created by the data team. If a question has been asked and answered in the past, Secoda will automatically index that question and answer to help the end-user self serve their answer and to make sure that the data team never has to answer the data request twice. Secoda also indexes all other data knowledge (data catalogue, metadata, docs, queries) to help anyone self-service in case there isn’t an answer in Secoda. This type of automation can help a small, underfunded data team start to amplify their impact across the organization. 

5. Measure and improve the data requests workflow

Lastly, you can’t improve anything you don’t measure. Taking the time to measure what your users are asking, which tables are used the most, and who is the most influential user in your organization is a great way to automate more common questions. Secoda does this automatically by showing you all the analytics related to your data knowledge from one place.

To all data teams who have been thinking about improving their operations, we hope this is an easier way to manage all the chaos and disjointed information that exists across the organization. Managing data requests in a scalable way is one way to help set up your team for success as you scale!