SC
Strategy Map›Layer 4 · Product Strategy›Grows as you use it+ from outside the named canon

Data Gravity

What is Data Gravity?

Data gravity is the pull a product gains when a customer's shared record lives inside it and new products, teams or uses read and write that same record instead of a copy. Every new use starts with the data already in place, which makes the product the natural home for the next one and hard to leave.

Also calledsingle source of truthsystem of record

How Data Gravity works

Each new use starts with data already in place, so adding it costs less than starting elsewhere, and every use leaves the record more complete. Workday launched Financials in July 2007 on the same foundation as its HR product; Yale added Financials in 2017 on the HR data it had kept there since 2015. Procore grew from four products in 2017 to 13 in 2020, all on one project record running from bid to payment. Leaving either company means moving the whole record.

HR records in WorkdayPeople and approval data on one foundation
Financials on that foundationLaunched July 2007
Yale adds financeIn 2017, on its 2015 HR data
One record, many usesFinance reads the people and approvals HR keeps
How Workday extended one shared record from HR into finance, so a new product started with the customer's data already there.

How companies won with Data Gravity

From the win chains in Strategy Canon case studies: what the winner did, and what its rival did at the same step.

vs. Cylance

CrowdStrike

Every endpoint's activity in one cloud graph

What it did
Threat Graph held all customers' telemetry; 'step one was to get data', unlike prevention-first rivals (Kurtz)
Cylance
Detection and response added May 2017 as CylanceOPTICS, which kept data on the endpoint 'without requiring cloud connectivity'
Read the CrowdStrike case study →
vs. ASCO CancerLinQ · 2013–18

Flatiron Health

Own the oncology record, not a layer over it

What it did
About 20 centers by 2014, and Genentech said too few patients
ASCO CancerLinQ
Pulled data from practices' existing systems: eight record systems and 60 ways of recording one blood count in the 2013 prototype
Read the Flatiron Health case study →
vs. PlanGrid · 2016–25

Procore Technologies

One project record from bid to payment

What it did
Four products (2017) to 13 (2020); Honest Buildings financials (2019), Levelset lien management for $500M (2021), Procore Pay from invoice to payment (2023)
PlanGrid
Stayed a drawings and field tool; 'hit a wall in terms of TAM in SMB and midmarket' against enterprise-ready rivals (Young, 2023)
Read the Procore Technologies case study →
vs. Amazon Redshift · 2018-24

Snowflake

Switching costs

What it did
Consolidated analytics, logs and pipelines on one platform (Capital One 2017-20, DoorDash 2021)
Amazon Redshift
Matched the architecture late (RA3 2019, data sharing 2020, Serverless 2022)
Read the Snowflake case study →

Also tagged: HipChat · Klaviyo · Rubrik · Slack · Workday

Data Gravity vs. Switching costs and Multi-Product

What it isHow to tell it apartExample
Data GravityThe customer's shared record lives in the product, and new uses build on itA second product or team reads data another created, with no exportYale adding Workday Financials onto its HR data
Switching costsAny value a customer expects to lose by changing to an alternativeAn outcome with many causes; data gravity is one, through a shared recordSalesforce with years of customizations
Multi-ProductFurther products for adjacent jobs, sold to the same customersCheck whether the products share one record or keep separate dataHubSpot adding CRM, sales and service products

How to tell if a company has it

  1. A customer record lives in the productName the data, history or documents. A shared login, interface, contract or brand doesn't count.
  2. A second use builds on itOne product, user or team reads or writes data another created, without exporting or re-entering it.
  3. Dated evidence of the shared useA launch, customer adoption or filing that shows when the second use began working on the same record.
  4. Products share one data storeAcquired products left on separate stacks, or ones that need data moved between them, fail this check.

Who names it

AuthorWhat they call it / where it appears
Dave McCrory +Data gravity
Dave Yuan +Data gravity: a single source of truth