Google Sheets

What Google Sheets is

Google Sheets is Google’s web-oriented spreadsheet application. It organises information in cells and combines formulas, tables, charts and collaboration. It is used with a Google account and forms part of Google Workspace, whose editions add administration capabilities and features according to the plan.

Its distinguishing approach is working on a shared version rather than exchanging independent copies. This supports budgets, schedules, analysis and coordinated lists, while making it important to agree what each row represents and who may change the structure. In OSH it can provide a coordination and data-preparation space, with original records and decisions requiring professional review clearly identified.

Reference documentation: Google Workspace: Google Sheets.

Origins and major milestones

Google Spreadsheets launched as a limited test on 6 June 2006. Jonathan Rochelle’s announcement describes the team’s origins at 2Web Technologies and its aim of sharing information in real time. Collaboration was therefore central to the proposition from the outset.

Google acquired DocVerse in 2010 and Quickoffice in 2012 as it developed collaboration and Office-document capabilities. In April 2014, it announced standalone Sheets applications for Android and iOS. Development was not simply about adding formulas: it also expanded how files could be accessed and which devices could be used.

The historical Explore assistant was introduced in 2016 and should not be confused with current Gemini features. Macro recording arrived in April 2018, and tables were announced in May 2024 to give data a more explicit structure. Some earlier mechanisms disappeared: Google Cloud Connect was retired in April 2013. Preserving this distinction avoids following obsolete instructions for extensions or services outside the current workflow.

Reference documentation: Google Blog (2006): It is nice to share · Google Drive Blog (2010): Google Docs welcomes DocVerse · Google Blog (2012): Google and Quickoffice · Google Drive Blog (2014): New mobile apps for Docs and Sheets · Google Workspace Updates (2018): Think macro · Google Workspace Updates (2024): Tables in Google Sheets · Google Workspace Updates (2016): Explore in Docs, Sheets and Slides · Google Workspace Updates (2013): Cloud Connect shut down.

A shared spreadsheet with working views

Google Sheets organises data in shared spreadsheets. Its value extends beyond adding columns: it can provide a common place where several people maintain a register with agreed definitions. For reliable work, each row should represent the same kind of item and headings should explain what is recorded, its unit and its date.

Ordinary filters affect the shared presentation; filter views allow people to work with different selections. This is useful when each site reviews its outstanding items without changing everyone else’s view. Filtering changes what is displayed, not who is authorised to access the data. A convenient view must not be mistaken for a permissions boundary. Decide whether a colleague needs access to the underlying register or only to a summary of it.

Reference documentation: Google Docs Editors Help — Sort and filter your data.

Cells, formulas, tables and visualisation

A cell contains a value or formula; formatting determines its appearance. Sheets provides functions for calculations, dates, text, lookups and queries. A sum calculates a quantity, a count counts records, and QUERY can express selection and aggregation over a dataset. Define the question before selecting a function: counting training rows is not the same as counting distinct trained people.

Tables introduce an identifiable structure and column-organisation options. Charts and pivot tables summarise records differently without turning them into new facts. A line can show change, a bar compare categories and a table retain detail. Percentages require an explicit denominator, and filters must be interpreted alongside the period.

Dates are represented as serial numbers originating on 30 December 1899; the fractional part represents part of a day. The serial does not itself contain a time-zone offset. When exchanging historical dates or timestamps, check a known example, regional formatting and the meaning of an empty cell. Changing its appearance to a date does not correct a value imported incorrectly.

Reference documentation: Google Workspace Updates (2024): Tables in Google Sheets · Google Docs Editors Help: Function list · Google Sheets API: Date and number formats.

Collaboration, version history and offline work

Sharing a file creates a common workspace, with permissions suited to each person’s role. Comments support discussion or review requests; they should not replace a status field when the process needs summarising. Version history lets users inspect changes. Check approval separately.

Stable coordination means separating people who contribute information from those who change headers, formulas and catalogues. Keep an identified reference for the agreed plan and compare subsequent changes. Before reusing older data, check which recent contributions would disappear.

Offline work requires preparation in a compatible environment and making the file available for that use. When connectivity returns, check synchronisation and other collaborators’ changes. For a preventive visit with unreliable coverage, testing the actual device and file beforehand is more useful than assuming every link will remain available.

Reference documentation: Google Docs Editors Help: Find what changed in a file · Google Docs Editors Help: Work offline · Google Workspace: Google Sheets.

Automation, macros and extensions

Apps Script extends Sheets through JavaScript and integrates Google services. A custom function can be used inside a sheet, while a script can execute a sequence of operations under its authorisation conditions. The Macro Recorder generates a script bound to the file from recorded actions, which can subsequently be reviewed and changed.

Before automating a task, define its inputs, expected output and what happens if it runs twice. A routine that appends rows, for example, needs a key or a check to avoid duplicates on repeated runs. Code that worked with three rows may need adjustments for empty fields, structural changes or execution limits.

Third-party add-ons provide further possibilities but have their own permissions and maintenance. Integrating a public source, including an encyclopaedic table, does not guarantee accuracy or a permanently stable structure. OSH analysis should retain a dated copy of the inputs that explain the result rather than depending on a changing external page.

Reference documentation: Google for Developers: Apps Script overview · Google for Developers: Custom functions in Sheets · Google Docs Editors Help: Automate tasks in Sheets.

Connected sources and integration with other systems

IMPORTRANGE brings in a range from another spreadsheet with the appropriate access. A chain of connected sheets creates dependencies: a source-structure change may leave the result incomplete or misaligned. Importing a table and creating a live link are different decisions, particularly when a closed-period report must remain reproducible.

Connected Sheets supports analysis of BigQuery data through the Sheets interface. This does not turn the entire BigQuery dataset into local cells or remove source-system permissions and query costs. Define who can refresh, which query runs and how the last valid result is identified.

For a preventive record, input might originate in a form and subsequently feed a report. Access to the form, spreadsheet and report needs separate review. A technical connection between them does not establish who needs individual responses and who needs only a summary.

Reference documentation: Google Cloud: Use Connected Sheets · Google Docs Editors Help: IMPORTRANGE.

Gemini and assisted analysis

Current Gemini in Sheets documentation describes assistance with tables, formulas, charts and analysis, subject to an eligible plan and feature availability. A natural-language instruction can speed up preparation, but it should define what counts as a record, how to treat missing values and which period to use. Without these decisions, a fluent response can conceal assumptions different from those intended.

Simple ML for Sheets, introduced by the TensorFlow team in December 2022, is a machine-learning add-on distinct from Gemini. Training or applying a predictive model does not explain why a phenomenon occurs. In OSH, such output would require review of the data, population, errors and practical usefulness before informing decisions that affect people.

Reference documentation: Google Docs Editors Help: Collaborate with Gemini in Sheets · TensorFlow Blog (2022): Introducing Simple ML for Sheets.

Formats, limits and portability

Sheets can work with Excel files and exchange formats such as CSV. Editing an Office file while retaining its format differs from converting it into a native Google spreadsheet. Compatibility checks should cover formulas, dates, charts, permissions and automation; opening a file does not prove equivalent behaviour.

Google documentation checked in November 2026 states up to 20 million cells or 100 MB for spreadsheets created in or converted to Sheets. This published limit is not a recommended performance target: complex formulas, chained imports and formatting rules can make a much smaller file difficult to use. It should not be confused with the volume that can be queried through Connected Sheets.

Exporting data is only part of preserving or migrating a process. Calculation rules, scripts, relationships between files and an explanation of permissions are also needed. A CSV retains tabular values but does not itself reproduce the collaborative system that generated them.

Reference documentation: Google Drive Help: Files that can be stored · Google Drive Help: Work with Microsoft Office files.

Preventive work that can fit Sheets

• Preparing a follow-up meeting: collect progress and questions against identifiable action rows that can be reconciled with original evidence.
• Coordinating activity across sites: maintain proposed dates, owners and dependencies for a safety campaign. The register supports coordination; it does not prove that an activity took place.
• Reviewing response quality: identify incomplete fields or inconsistent categories before reporting, rather than letting each person reclassify records differently.

These are use cases proposed by Sabentis. Before adopting them, agree who owns the structure, what constitutes closure and how disagreements between contributors will be resolved. Keep the criteria available alongside the register so new contributors can use the same definitions.

Worked example: planning a campaign across sites

Hypothetical example. Five sites are preparing short awareness sessions and need to know which materials are ready, who will deliver each session and what still needs confirmation.

1. Create a register containing session identifier, site, planned date, owner, material version and status. Separate proposed dates from actual delivery dates.
2. Define a short status vocabulary, such as awaiting materials, awaiting scheduling and confirmed. Document the evidence required to move between states.
3. Create site-specific filter views and a global exceptions view. Check that nobody interprets a filtered view as the entire campaign.
4. Protect calculation columns to reduce accidental editing, without describing this as confidentiality protection.
5. Review a sample with the owners, resolve duplicates and retain a reference copy of the approved plan. After each session, add actual delivery information and a link to authorised evidence.

The acceptable output is a coherent, traceable schedule. A session confirmed in a diary and a session actually delivered are different events and should remain distinct in the register and every summary.

As an additional check, if the plan contains 25 sessions, with 18 delivered, four postponed and three unconfirmed, these categories should reconcile to the 25 identified sessions. Counting only the 22 with responses would use a different population. A single-site view should not supply the overall denominator when calculating delivery percentages.

Reference documentation: Google Docs Editors Help — Sort and filter your data · Google Docs Editors Help — Protect, hide and edit sheets.

Joint editing, confidentiality and interpretation

Google warns that protecting a worksheet or range should not be used as a security measure: collaborators may still have ways to copy or export information. Hiding a tab should not be treated as a confidentiality barrier either. Share only with appropriate recipients and review the file’s access permissions.

A collaborative register may have simultaneous changes to statuses, formulas and definitions. Decide who can alter the structure and how changes are communicated. Before presenting figures, check hidden rows, filters, date formats and duplicates. Do not replace a missing response with a negative confirmation or a zero; it needs clarification. Keep a distinction between an unresolved question and an approved operational instruction.

Reference documentation: Google Docs Editors Help — Protect, hide and edit sheets.

When to choose Sheets and when to use another tool

Sheets fits situations where direct collaboration and a relatively simple register matter more than complex approval or access workflows. Choose a different structure when the process requires strict separation of which records each person may see, or controls unavailable in the current file configuration.

Compare Excel for analysis and data preparation in a Microsoft environment. To collect responses without exposing the register, consider Google Forms. Review ownership and permissions for the form and spreadsheet separately.

Frequently asked questions about Sheets in OSH

Does a filter view hide data from other sites?

It is not a confidentiality control. It organises displayed rows but does not replace file-access permissions.

Should health information be entered in a shared sheet?

First establish whether that processing is necessary and authorised. Use the minimum required detail in an appropriate environment; easy sharing does not justify collecting sensitive information.

How can the summary be checked for completeness?

Reconcile record counts with the source and review filters, duplicates and missing values. Retain the reporting date and applied criteria.

Reference documentation: Google Docs Editors Help — Sort and filter your data · Google Docs Editors Help — Protect, hide and edit sheets.