Why Start-up CEOs no longer require a full-blown data team

Introduction
Over the last few years, the Modern Data Stack has made it easier than ever to feel at the forefront of Data and Analytics.

However, the Modern Data Stack hasn't made life any cheaper for people setting out.

In fact, the pattern for many start-ups has been for the CEO to boldly declare the company is becoming "data-driven", and swiftly look to hire a Head of Data or Lead Data Engineer.

This has never really been a good move. Now, it's an inexcusable one.

Problems facing a "one-man" data team
Being a one-man data team in an organisation of 30–100 is extremely tricky. In fact, it's arguably impossible.

Here are the responsibilities of a one-person data team:

• Choose what architecture to use
• Decide what Data Products to Build
• Maintain Architecture and fixed pipelines
• Write code to ingest, transform and surface data
• Build Dashboards
• Advocate Internally
• Educate business users on new data sources

This is simply too much. At a cost of perhaps £70,000, it's also an enormous investment for a business looking to wet its toes with data.

Furthermore, it often leads to chaos.

Individuals making significant decisions on their own do so in a way that impacts the business forever.

A single data engineer may choose to ingest data using a home-grown application rather than a SAAS tool, thereby incurring years of tech debt in months. Conversely, they may choose a SAAS tool but be unaware of its limitations and "quirks", — resulting in bills that resemble years instead of months.

The New Playbook
Well-informed CEOs are increasingly staying away from this.

They're adopting a new playbook, which looks more like the following:

Consultant Set-up
CEOs are preferring to get small or even single-person consultancies to set up their Data Stack.

This has two advantages.

The first is that those doing the implementation have done this dozens of times before. As long as the information around guardrails is applied to elements of the Modern Data Stack, it's hard for the CEO to be left with something that isn't fit for purpose or balloons in cost.

The second is risk — by spending between £10k to £30k on a consultant or part-time work, the ramp-up time is lower and the risk of the project failing is lower.

This, therefore, represents a much more economically favourable option. It is also more consultative and involves business stakeholders (other than the data team) from day 1—something absolutely fundamental when considering how best to implement a data-driven culture.

How Orchestra alleaviates integration pain
This approach has been facilitated by two important trends in data.

The first is the data industry's age.

Data is a relatively new industry and, as such, hasn't always benefited from a large corpus of experienced consultants who are willing to provide helpful advice at an affordable price.

As it matures, there are increasingly more and more options. For example, Ben Rogojan (the Seattle Data Guy), Steinert Analytics or Hypermetric. Increasingly, many Data Engineers are offering consultative services.

The second trend is increasing interoperability and focus on git-controlled, GUI-driven tools like Orchestra.

The ability to easily spin-up an environment with end-to-end orchestration and obsverability has historically required at least a single person to do.

Workflow Orchestration tools like Airflow, Prefect, Dagster and Mage all require a person to maintain them. Every time a new tool is added, new code must be written.

Gaining end-to-end visibility into how well the data organisation is functioning is also highly dependent on additional pieces of observability software that are technically-minded and not business-focussed, like Monte Carlo, Datafold or Metaplane.

These tools are also extremely expensive.

Platforms like Orchestra allow single-person data teams to quickly choose whatever pieces of software they want while retaining visibility and end-to-end orchestration without shelling out hundreds of thousands of dollars and committing to hiring a data engineer with platform engineering experience.

Summary
CEOs are no longer hiring a single-person data team as their first port of call when setting up a Data and Analytics Function.

Instead, they're focussing on leveraging small and boutique data consultancies and using components of a Modern Data Stack.

This can be complemented by some existing resource with some technical capability e.g. a CTO or Software Engineer.

This is rapidly accelerating organisations' move to the cloud and significantly, rapidly increasing the rate at which businesses get value from their data.