Data

Establishing Data Analytics at a Start-Up: 5 Lessons Learned

By Ausra Puniskyte

Having recently reached my one-year milestone at a start-up, I found myself reflecting on the journey of establishing analytics within the company.

So, if you’re looking to set up analytics at a start-up, here are five essential steps to consider.

1. Understanding Your Purpose

Starting off strong with something that might seem obvious, but it’s not necessarily your job description: understanding why analytics exists at the company in the first place. Every analytics team has a specific purpose and is responsible for particular use cases. It’s crucial to clearly define the responsibilities of the analyst and the ownership of data flows.

2. Data Pipeline Design

One of the most important steps is to understand the data pipeline. This includes considering the datasets currently available, as well as any potential ones that may become available at a future date. Here are the questions to ask:

  • How does the data reach you?
  • What actions do you take with the data?
  • Are there any steps required to obtain the data from its source?

The answers to these questions help construct a data pipeline. A typical industry data pipeline consists of the following steps: data collection, data ingestion, data storage, data analysis and modeling, and data consumption. The complexity pipelines may vary based on how established your business is.

  • Data Collection: Identify sources where data is gathered, such as applications already in use within your organization. For example, retrieving data via APIs.
  • Data Ingestion: Utilize tools that automate the process of acquiring data from the specified sources like Fivetran.
  • Data Storage: Determine where all data will be stored. Options include dedicated centralized database storage solutions like Salesforce, Snowflake, etc.
  • Data Analysis and Modeling: Choose methods to aggregate, model, and compute the data for analyses.
  • Data Consumption: Consume the findings and share them with the team.

The final version of the design should resemble the infographic below. The specifics of these steps may vary based on the company’s operations and financial position, and individuals may choose to exclude certain steps if they are not necessary.

3. Tech Stack

The choice of technology stack is dependent on the proposed data pipeline and the resources available to the company. For each pipeline step, it is important to clearly understand what tool will execute the required tasks. While some tools come at a cost, there are many free options to consider. An analyst’s key focus should be where to store data, how to analyze it, and how to share your insights with the team.

Data Storage

At a small start-up, obtaining and managing a data warehouse may be particularly time-consuming and costly – this, of course, needs to be weighed against the size of the datasets to be dealt with. Some of the best data storage options I discovered during my research are Salesforce, Amazon Redshift, Snowflake, and GCP. Don’t underestimate the power of spreadsheets (specifically for very small start-ups).

Data Analysis and Modelling

What are you using to process, aggregate, and analyze the data? The most common tools to perform this (again depending on your organization’s provisions) are SQL and Python or data exploration tools like Tableau Prep and dbt.

To decide on an analytics tool to use, there should be three things to consider:

  1. A tool you are familiar with
  2. Something easy to use (by you or your teammates)
  3. And something that will give you room to upskill and grow

Free Data Visualisation Tools to Consider

  1. Tableau Public Desktop
  2. Looker Studio
  3. Datawrapper
  4. Google Sheets
  5. Power BI (free with Microsoft Business Package)
  6. Matplotlib for Python

4. Own, Advise, and Update

This is about taking pride in the work. As the owner of the analytics function, the responsibility for many data-related aspects falls on the analytics team. It’s crucial to understand the purpose of the function clearly.

The role extends beyond managing analytics – acting as an advisor to teams across the company and being the subject expert in the field are crucial too. In a start-up, where teams are typically small, this expertise is even more critical. Instead of waiting for problems to come to you, seeking out ways to help the business perform is the right decision to make.

It’s common for teams to lose interest in reporting, but it’s the responsibility of an analyst to keep the enthusiasm alive. This can be done by providing regular updates via Slack or dedicating time during company-wide meetings.

5. Knowledge

Keeping up-to-date with the newest affairs in the industry, learning a new skill, or even trying something new are ways to continue the learning curve. Whenever I find friction in my area of work, I may take a step back to see if I could do something different than usual.

Most recently, I found myself learning a few new applications. I’ve joined the world of dbt to streamline my SQL flows more efficiently. In addition, I decided to explore a new visualization tool. I’ve always been a big advocate for Tableau but recently discovered the greatness of Looker and have been invested in learning since.

Summary

Establishing analytics at a start-up comes with unique challenges, but it’s crucial for driving informed decision-making. By clearly defining your purpose, creating a solid data pipeline, choosing the right tech tools, and taking ownership of your role, you can build a strong foundation for your analytics efforts.

Any thoughts? Share them in the comments below!

The Author

Ausra Puniskyte

Ausra is a data-driven analyst with extensive experience using Tableau and Alteryx to analyze and visualize data in the most impactful way.

Leave a Reply