Assessing your data maturity

Introduction

Modern businesses can be thought of as machines with thousands of sensors attached to them. Each sensor captures information from the world that operators can and should use to fine-tune the machine.

How well an operator reads this information and makes sense of it is not a solved problem. Getting foundational Data Operations weaved into the fabric of your company can often dictate how quickly you bring value to your customers and shareholders.

This is made harder by the fact that an entrepreneur’s priorities in the early days are disparate, and driving quality decisions using data may often fall by the wayside. On the flipside, going too deep and investing too much, too early, may mean that running the business will suffer and returns will be delayed. Striking that balance is tricky.

This guide is meant to provide a broad set of opinions that we hold on what the data maturity curve looks like for typical organisations as they grow. We’ve kept it general and it’s likely that your data set up will differ (as it should!) as it should be tailored towards your operating philosophy, organisational culture and business model.

Early Stage

Company Context

You may be a new startup, or a traditional enterprise just starting out on your data journey. The company has just found product-market fit (PMF) and have validated against your customer - now your company is doubling down on executing well: Win customers, grow revenue by selling to bigger and better contracts.

Data Operations at your Company

At this point, there are likely <10 weekly users who need to view reliable information for decision making.  The organisation might have survived to-date by querying against production databases or using predefined reports in your ERP or other SaaS tools.  You probably need to deliver on a specific ask/use case by your leaders in a somewhat reliable/stable way.  An example is revenue numbers for investors, or for north star metrics such as daily/weekly/monthly active users.  You likely don’t have a Data team to speak of; just some inspired engineers/ analysts who double-hat on ad-hoc reporting needs.

Transition point

You may have realized that running analysis off your production database is not the best idea, that one person in the organization is becoming a bottleneck/key-man-risk for mission-critical business information, or that you're running into the limits of Microsoft Excel.  This is the moment you need to decide on a Data Strategy.  Think through the levers of your business and which ones need to be augmented with top-notch data quality, how you’ll ensure robust capture and consumption.  The aim at this point is to set extensible foundations so that you can evolve your technology investments as the company matures, and avoiding locking yourself into a platform with high switching costs.

Call us biased perhaps, but we do think that bringing in external eyes at this point is a highly valuable exercise, as it will help you make 'directionally correct' early steps and set you up for efficient growth.

Mid Stage

Company Context

Your company has likely established itself in an important market niche, and there may now be a greater need to focus on driving performance in line with business priorities and against your competitors.  The difference between whether you survive or thrive in the next few years is your ability to execute and remain nimble amid changing circumstances.

Data Operations at your Company

From a usage standpoint, you might have <5 power users who look at info daily or perhaps <100 users who look at info monthly.  You may have 2-3 discrete basic use cases e.g. customer funnel, marketing attribution, revenue reporting, etc., including those which cut across oranisational silos.  The company has established central data storage infrastructure to enable this cross-functional analysis.  Users could be self-serving with off-the-shelf tools or be supported by a pool of 2-5 data specialists.  Questions up until this point are more commonly 'do we have this data?'

Transition Point

You increasingly start getting questions about data quality (freshness / trust).  Teams may be reporting the same core concept but presenting different numbers to company leadership.  Resolving differences in core concepts (e.g. a user from your click stream/website data vs the same user in your application data vs in your email remarketing tools) takes a long time and there is no clear owner to make these decisions.

At this stage, you have some discrete use cases you want to build solutions for. This is likely something that will help orient functions in your company (i.e. Product / Sales / Marketing / Ops) to some metrics that the company needs to execute well on, and be able to identify driver trees for.

You may start seeing a wedge amongst data producers (engineering/product teams) and data consumers (marketing/sales).

Establishing a Data Operating Model which outlines each actor’s roles and responsibilities is important, alongside tooling that helps to surface data defintions and manage information quality.

Late Stage

Company Context

Your company likely has several million in annual revenue, and is using an enterprise solution for Business Intelligence.  There is an established data engineering and/or analytics team(s) who manage towards service level agreements.  Automation and operational efficiency has been a key business priority, and you are starting on predictive (i.e., what do we forecast will happen) and prescriptive (i.e., if an outcome looks like it will occur, how do we mitigate it) analytics, potentially with some machine learning or statistical models to support this operation.  At this point, your data use case is highly contextual to the company you operate.

Data Operations at your Company

In terms of usage, you probably have 20+ power users who look at info daily, and ~ 300 users who look at info monthly.  You may have 5-8 use cases, some of which may be advanced e.g. A/B testing, dynamic active user calculations, calculating business hours excl public holidays between timestamps, etc.  Teams across the organisation have created hundreds of data models, dashboards and reports.  Data people are part of a matrixed organisational structure and may have both formal and dotted reporting lines.  There is some sensitivity on how to share figures across the org.

Transition point

Questions come up about dashboard or report performance, and data teams' time is often consumed with refactoring models to make them scalable and performant.  Costs may be a concern as unbridled growth of data users has led to duplicated or orphaned data assets.  As a result you may find that you need to establish controls over how data is consumed, i.e. establish data governance. Related to that, you may need to expose some metadata about data assets and models to users, and help them decide which datasets are most appropriate to use for self service.

Analysts working across the organisation often have different approaches but don't always share best practices with each other, particularly if they have formal reporting lines into different business units (e.g., finance vs sales analysts).  There might be some level of operational dependence (as opposed to strictly reporting dependence) on your data models.  This could be in the form of embedded dashboards, loading from the data warehouse into sales and marketing tools (i.e., reverse ETL) to enable marketing campaigns, ML models, etc.  You likely need to make commitments to ambitious SLAs e.g. 99.9% data availability before 9 am in the morning, especially for complex transformation pipelines.

At this point, a scalable data platform that enables discovery of both raw data and analysis, reinforces trust in data, and supports data asset administration and governance starts to become critical.

Conclusion

There are many pathways through these maturity stages.  If you’re starting out, or have been doing this with some major outstanding questions, know that you’re not alone.  You also don’t need to suffer alone, reach out to us here for an introductory chat!  We have run data maturity assessments for businesses across all these stages.  Our team will conduct a structured survey of your organisation, supported by targeted stakeholder interviews and deliver recommendations on the people, process and infrastructure investments that match the way you currently use data or want to into the future.

Cover image by freepik