Starting an Analytics Org From Scratch — Lessons From a Decade at DoorDash

Starting an Analytics Org From Scratch — Lessons From a Decade at DoorDash

DoorDash VP Jessica Lachs has grown the company’s analytics team from a band of scrappy generalists to a highly specialized 300-person org. Here, she spells out how founders can lay the groundwork for an analytics function at their startup and pick the right first hires for the team.

ANALYTICS 101: FIGURING OUT THE “WHAT” FOR YOUR ORGANIZATION

Focus on your needs, not nomenclature

The ambiguity of the term “analytics” can create confusion. Is analytics different from data science? Business intelligence? Product analytics? Machine learning? BizOps? The answers will differ depending on who you ask — it varies from company to company.

Instead of questioning the semantics, decide on the outcomes you want first.

Generally speaking, the analytics team's role is a function of a company's data-centricity. Some companies don’t prioritize data in decision-making, perhaps because they don’t have reliable data yet at their size or stage — or because leaders’ past experience with analytics was limited to business reporting. Other companies, like DoorDash, prioritize the use of data in decision-making, giving analytics a broader remit and a seat at every table where a business decision is being made.

But even within the same company, the role of analytics can change over time as the business matures — as it did at DoorDash. When a company is in its early stages, there’s less data to analyze, and decisions are often more straightforward. Making decisions by gut instinct can lead to good outcomes for some time, but it doesn’t scale.

The teams you need will change based on the types of problems you have, the amount of data you have, and how you plan to use data for decision-making.

The three C’s of an analytics function

There are three core elements of analytics: creating the data you need, curating that data for usability, and helping users accurately interpret and consume that data to make decisions.

Analytics sits on the spectrum of data creation, curation, and consumption.

Data creation

Having reliable data is the foundation of any analytics org. There are many types of data, but at DoorDash, we focused on two types:

  1. Transactional data: Purchases and related attributes, including time, location, price, payment method, and discounts of that purchase
  2. Behavioral data: What the user does in the mobile or web app

These are some other types of data you might create or collect:

Data curation

Data curation is developing and maintaining data models to democratize data through easy access and use. This includes data munging, which is the process of cleaning, parsing, and validating data before it’s ready to be used.

Early at DoorDash, queries were slow to run and dashboards frequently timed out. This wasn’t surprising because getting an output required complicated underlying SQL code across many data tables.

Data consumption

While the analytics function at DoorDash has evolved, the mission hasn’t changed: to continually improve the business through scientific and data-driven insights.

With the existing data in an accessible form, ready for consumption, there’s a lot the analytics team can do to understand the key business levers, identify opportunities to improve, and measure the impact of the product or business decisions.

For example, at DoorDash, we believed that speed was important to consumers. We had robust data on how long a delivery took
from when a consumer placed the order to when it was delivered. We identified an opportunity to increase the accuracy of our ETAs, improving the consumer experience and driving meaningful growth.

Structuring data consumption at each stage of a startup

Let’s dig deeper into the analytics function and how data consumption translates into business impact at different company stages. There are three focus areas:

  1. Decision tools and measurement
  2. Deep dives and analysis
  3. Experimentation

The level of complexity and sophistication in each area will change as a company matures.

The 0 - 1 stage: < 100 people, Seed or Series A

At this stage, analytics should focus on defining and tracking the core metrics to measure product-market fit.

For DoorDash in the early days, we tracked these metrics:

Early-stage companies: 100 - 400 people, Series B or C

Startups in this stage will need to decide what kind of analytics function they want. There are two paths: analytics as a measurement function or a team that uses data to drive business impact.

Later-stage companies: > 400 people, Series D or post-IPO

Companies that want to use data to drive business impact must balance the need for immediate access to data to inform decision-making with the investment in the long-term scalability of the data and analytics platform.

HIRING FOR ANALYTICS: FINDING THE RIGHT “WHO” FOR YOUR EARLY TEAM

Because the role of analytics in the early stages is broad, I recommend starting with generalists for the earliest hires. A great generalist can take on most challenges you throw at them.

Hard and soft skills to look for

Here are the skills I assess candidates for when hiring for a generalist:

Hard skills

Soft skills

Hire internally if you can

But it won’t always be necessary to hire externally for this team. Depending on the complexity of the problems, someone from an internal team may be able to fill the role.

Pick mid-level ICs early on

For an early hire, you want to find someone who knows how to do the work and is excited to get their hands dirty.

Add specialists as you grow

As your analytics organization matures, you may find that you need to expand your hiring to include more specialists.

Sample profiles of early analytics hires

Top talent can come from various places and backgrounds. Here are some profiles that I have had success hiring over the years:

WRAPPING UP: AN ANALYTICS READINESS CHECKLIST

An analytics team can add value at any stage by explaining what’s happening in the business, determining whether a company should take action based on current performance, and uncovering new opportunities for growth and profitability. Here are some helpful questions to help you figure out if you’re ready to hire analytics team members: