Growth ops
How growth teams measure, test and prioritise.
75 tactics
'How did you hear about us?' self-reported attribution
Ask new signups or buyers where they heard about you, ideally in a free-text field, and use the answers next to software attribution to see channels that tracking misses.
WorkingSecondary20251 example · 3 sourcesGrowth ops · Analytics✓ readA/A tests to validate the setup
Run a test where both groups get the same experience to prove the experimentation pipeline works.
WorkingVendor claimno examples0 examples · 2 sourcesGrowth ops · Paid✓ readAgent-built multi-touch attribution
Use an AI agent to join CRM, warehouse and call data and separate last-touch source from earlier influences.
WorkingSecondary20261 example · 1 sourceGrowth ops · Paid✓ readAI agent for ad operations
Replace manual or agency campaign set-up with an AI skill that builds and optimises campaigns across ad platforms.
WorkingSecondary20261 example · 1 sourceGrowth ops · Analytics✓ readAI-generated experiment reports
Use an analytics tool's AI to write the structured experiment summary and check results, then regenerate as data grows.
WorkingVendor claimno examples0 examples · 2 sourcesGrowth ops · Process✓ readAI-native growth engineering team
Staff growth with engineers who own an outcome end to end and use AI coding and analysis tools to run experiments at high volume.
WorkingSecondary20261 example · 2 sourcesGrowth ops · Analytics✓ readAnalytics AI agents with set autonomy
Use AI agents inside the analytics stack to answer questions, build dashboards and watch session replays, with modes controlling how far they act alone.
WorkingVendor claimno examples0 examples · 2 sourcesGrowth ops · Outbound✓ readAuthenticate sending domain (SPF, DKIM, DMARC)
Publish SPF, DKIM and a DMARC policy for every domain used for outbound, now a baseline rule at Gmail, Yahoo and Outlook for large senders.
RestrictedPrimary2024 · 20253 examples · 3 sourcesGrowth ops · Partners✓ readBuild one MCP server for many AI clients
Treat MCP as the shared integration layer so one server serves Claude, ChatGPT and developer tools.
WorkingPrimaryno examples0 examples · 2 sourcesGrowth ops · Paid✓ readBulk ad variants with Claude Code sub-agents
Use sub-agents to generate large batches of ad copy variations within platform limits, replacing manual copy-and-paste.
WorkingPrimary20251 example · 2 sourcesGrowth ops · Outbound✓ readCAN-SPAM-compliant cold email
Make every cold email meet US CAN-SPAM rules: honest headers, postal address, working opt-out honored in 10 business days.
RestrictedPrimary20231 example · 1 sourceGrowth ops · Billing✓ readCancel as easy as sign-up (click-to-cancel status)
Design subscription cancellation to be as simple as enrollment; track the shifting US federal and state rules.
RestrictedSecondary20251 example · 2 sourcesGrowth ops · Analytics✓ readCheck for novelty and primacy effects
Account for effects that fade (novelty) or grow (primacy) over time before treating early lifts as permanent.
WorkingPrimaryno examples0 examples · 2 sourcesGrowth ops · Analytics✓ readCheck for Simpson's paradox and short-term bias
Review known experiment pitfalls (Simpson's paradox, short-term versus long-term effects) before declaring a winner.
WorkingPrimaryno examples0 examples · 2 sourcesGrowth ops · Analytics✓ readCohort analysis for retention and growth
Group users by shared date or behavior and compare how each group retains over time to find what drives lasting use.
WorkingVendor claimno examples0 examples · 2 sourcesGrowth ops · Analytics✓ readConsistent UTM tagging convention
Tag every link you control with source, medium and campaign parameters following one written, case-consistent naming scheme, so all campaigns roll up cleanly in analytics.
WorkingPrimaryno examples0 examples · 1 sourceGrowth ops · Analytics✓ readContextual bandits to personalize creative
Use a bandit that also takes user attributes into account to choose which version of an image, headline or offer each user sees, learning online instead of waiting for a batch model and an A/B test.
WorkingSecondary1 example · 2 sourcesGrowth ops · Analytics✓ readCUPED variance reduction
Adjust each user's metric with their pre-experiment data to cut noise and reach conclusions faster.
WorkingVendor claimno examples0 examples · 2 sourcesGrowth ops · Analytics✓ readData-driven attribution in ad and analytics tools
Let the analytics or ad platform's model assign conversion credit to each touchpoint based on observed paths, instead of fixed rules, and feed that credit into automated bidding.
WorkingVendor claim20231 example · 1 sourceGrowth ops · Process✓ readDemocratized experimentation with safeguards
Make experiments cheap and safe enough that anyone can run them, with automated safeguards and the right to stop harmful tests.
WorkingPrimary20171 example · 2 sourcesGrowth ops · Analytics✓ readDiscount summed A/B winners (winner's curse)
Do not add up the lifts of all winning tests; the sum overstates the true effect, so measure cumulative impact with a holdout or an adjusted method.
WorkingVendor claimno examples0 examples · 2 sourcesGrowth ops · Process✓ readEarly growth team: prove value, then scale
Start with one or a few people who cover both business thinking and implementation, win visibly, and scale to 10 to 20 once results justify it.
WorkingPrimaryno examples0 examples · 2 sourcesGrowth ops · Product✓ readEnd-to-end PLG audit
Audit the whole self-serve journey step by step (trial, activation, paywall) to find quick wins in activation and conversion.
WorkingUnverifiedno examples0 examples · 2 sourcesGrowth ops · Billing✓ readEU withdrawal button
Provide a visible online 'withdraw from contract' function for distance contracts with a withdrawal right.
RestrictedSecondaryno examples0 examples · 3 sourcesGrowth ops · Analytics✓ readEvent taxonomy and naming convention
Define which events and properties you track, how they are named and how they fit together before instrumenting.
WorkingVendor claimno examples0 examples · 2 sourcesGrowth ops · Process✓ readExperiment doc and learning log
Write a short doc for every test and log the result, including failures, so learning compounds and is shared.
WorkingPrimaryno examples0 examples · 2 sourcesGrowth ops · Analytics✓ readFixed sample size, no early stopping
Decide sample size before an A/B test and avoid stopping at the first significant result.
WorkingSecondaryno examples0 examples · 2 sourcesGrowth ops · Analytics✓ readForces of progress analysis
Explain a buying decision through four forces: push, pull, anxiety and habit, and reduce the forces holding customers back.
WorkingPrimaryno examples0 examples · 2 sourcesGrowth ops · Process✓ readFour Fits growth audit
Audit growth by checking four fits that must hold together: market-product, product-channel, channel-model and model-market.
WorkingPrimaryno examples0 examples · 2 sourcesGrowth ops · Paid✓ readGeo holdout incrementality tests
Measure what ads truly add by turning them off, or on, in randomly chosen regions or audience holdouts.
WorkingPrimary20251 example · 5 sourcesGrowth ops · Analytics✓ readGrowth model and levers tree
Build a model of the variables that drive acquisition, retention, engagement and monetization, and use it to set goals and pick bets.
WorkingPrimaryno examples0 examples · 2 sourcesGrowth ops · Process✓ readGrowth pods matrixed to functions
Organize growth as pods that each combine product, engineering, design and data, matrixed to their home functions.
WorkingAnecdotal1 example · 2 sourcesGrowth ops · Process✓ readGTM engineering function
Create a role that builds automated revenue systems with AI, data enrichment and workflows, working like an internal product team for sales and marketing.
WorkingVendor claim20261 example · 3 sourcesGrowth ops · Analytics✓ readGuardrail metrics on every test
Monitor metrics you must not break (errors, speed, retention) alongside the primary metric of every experiment.
WorkingVendor claimno examples0 examples · 2 sourcesGrowth ops · Process✓ readHigh-tempo growth testing
Run many small, scored experiments every week from a standing backlog instead of a few big launches per quarter.
WorkingPrimary2016 · 20263 examples · 3 sourcesGrowth ops · Process✓ readICE scoring for test ideas
Rank experiment ideas by scoring each 1 to 10 on impact, confidence and ease, then work from the top.
WorkingPrimary20161 example · 2 sourcesGrowth ops · Process✓ readIndependent vs functional growth team
Choose between a dedicated growth org reporting to the top, or growth work embedded in functional teams under product.
WorkingSecondary20162 examples · 3 sourcesGrowth ops · Process✓ readInnovation over optimization in AI-era growth
In fast-moving markets, spend most growth effort on new product and distribution ideas rather than funnel tweaks.
WorkingAnecdotal20261 example · 2 sourcesGrowth ops · Analytics✓ readJourney map and bottleneck focus
Map the customer journey with data, find the step where most people drop off, and aim all effort at it.
WorkingSecondaryno examples0 examples · 2 sourcesGrowth ops · Analytics✓ readJTBD switch interviews
Interview people who recently switched to your product (or away) and walk forward through concrete events to learn why.
WorkingPrimaryno examples0 examples · 2 sourcesGrowth ops · Outbound✓ readKeep spam complaint rate under 0.3%
Monitor the complaint rate in Google Postmaster Tools and cap volume or fix targeting when it nears 0.3%.
RestrictedPrimary20242 examples · 2 sourcesGrowth ops · Marketplaces✓ readList an MCP connector in Claude's directory
Submit a remote MCP server to Anthropic's Connectors Directory so it can appear in Suggested Connectors across Claude products.
WorkingPrimaryno examples0 examples · 3 sourcesGrowth ops · Analytics✓ readLong-running product holdout groups
Keep a small share of users from new features for months to measure their combined effect.
WorkingVendor claim20261 example · 4 sourcesGrowth ops · Billing✓ readMerchant of record for tax handling
Use a merchant of record that legally resells your product and handles VAT and sales tax.
WorkingVendor claim20261 example · 1 sourceGrowth ops · Analytics✓ readMMM calibrated with lift experiments
Run an open-source marketing mix model and use lift or geo experiments as priors to keep its channel estimates honest.
WorkingVendor claimno examples0 examples · 4 sourcesGrowth ops · Analytics✓ readModel growth as closed loops, not funnels
Plan growth as loops where an output feeds the next input, instead of one-way funnels.
WorkingSecondaryno examples0 examples · 1 sourceGrowth ops · Analytics✓ readMulti-armed bandits for short-lived variants
For promotions, headlines and other short-lived content, let a bandit algorithm shift traffic to the best-performing variant during the test instead of holding a fixed A/B split.
WorkingVendor claim1 example · 3 sourcesGrowth ops · Billing✓ readNegative-option disclosure and informed consent
Disclose recurring-charge terms before taking billing details and obtain express consent.
RestrictedPrimary20251 example · 2 sourcesGrowth ops · Analytics✓ readNorth Star metric with input metrics
Pick one metric that captures the value customers get, then break it into a few input metrics that teams can move in their daily work.
WorkingVendor claimno examples0 examples · 2 sourcesGrowth ops · Outbound✓ readOne-click unsubscribe header on marketing mail
Add a working List-Unsubscribe header so recipients can opt out in one click; required for bulk marketing mail at Gmail and Yahoo.
RestrictedPrimary20242 examples · 2 sourcesGrowth ops · Process✓ readOne-to-two-week growth sprints
Apply short agile sprints to one learning goal, post experiments publicly, and review quickly.
WorkingSecondary1 example · 2 sourcesGrowth ops · Paid✓ readOpen-source marketing mix modeling
Use open-source MMM tools to estimate each channel's contribution and saturation without relying only on click attribution.
WorkingVendor claimno examples0 examples · 2 sourcesGrowth ops · Community✓ readOrbit model for community measurement
Measure a community by gravity (love times reach) and sort members into ambassadors, fans, users and observers instead of treating them as funnel leads.
WorkingSecondary20211 example · 2 sourcesGrowth ops · Product✓ readPick the network effect you can actually build
Choose which kind of network effect (physical, protocol, marketplace, data, tribal and others) fits and tune the product to reinforce it.
WorkingSecondaryno examples0 examples · 2 sourcesGrowth ops · Process✓ readPIE framework
Rank page or test ideas on potential for improvement, importance of the traffic, and ease of building the test.
Secondaryno examples0 examples · 1 sourceGrowth ops · Process✓ readPMF engine: high-expectation segment and split roadmap
Narrow to the users who would be very disappointed without the product, learn what they love and what blocks the rest, and split the roadmap between the two.
WorkingPrimary20171 example · 2 sourcesGrowth ops · Process✓ readPMF survey and the 40% test
Ask users how they would feel if they could no longer use the product and treat 40% answering 'very disappointed' as the PMF bar.
WorkingPrimary20171 example · 2 sourcesGrowth ops · Analytics✓ readPower analysis and MDE planning
Estimate how long a test must run to detect the smallest effect you care about before launching it.
WorkingVendor claimno examples0 examples · 2 sourcesGrowth ops · Process✓ readPXL binary test prioritisation
Prioritise tests by answering yes/no questions about each idea, so evidence rather than opinion drives the order.
WorkingPrimaryno examples0 examples · 1 sourceGrowth ops · Process✓ readRe-find PMF every three months (AI-era claim)
A claim from an AI-company growth leader that AI products must re-find product-market fit roughly every quarter and ship a 'minimum lovable' product.
WorkingUnverifiedno examples0 examples · 2 sourcesGrowth ops · Process✓ readRICE scoring
Score projects on reach, impact and confidence, then divide by effort to get impact per unit of work.
WorkingPrimary20181 example · 1 sourceGrowth ops · Analytics✓ readRules-based multi-touch attribution models
Split conversion credit across touchpoints with a fixed rule, such as all to the first click, evenly, more to recent touches, or most to the first and last touch.
RestrictedPrimary20231 example · 1 sourceGrowth ops · Process✓ readRun GTM engineering like a product team
Run internal GTM engineering with tickets, two-week sprints, version control and release notes, instead of ad hoc ops requests.
WorkingVendor claim20261 example · 2 sourcesGrowth ops · Analytics✓ readSample ratio mismatch check
Automatically test that the observed split of users between variants matches the planned split, and distrust results when it does not.
WorkingVendor claim20231 example · 3 sourcesGrowth ops · Paid✓ readSend CRM lead outcomes back to ad platforms
When a lead from an ad later becomes qualified or signs a contract, send that outcome (with the click ID and hashed contact data) back to the ad platform so bidding optimises for real pipeline, not form fills.
WorkingVendor claim1 example · 2 sourcesGrowth ops · Analytics✓ readSequential testing for early stopping
Use a sequential test design so you can stop early on a clear winner without breaking statistical validity.
WorkingSecondaryno examples0 examples · 2 sourcesGrowth ops · Paid✓ readServer-side conversion tracking (Conversions API)
Send conversion events from your server to Meta alongside the browser pixel to feed optimisation.
WorkingVendor claimno examples0 examples · 1 sourceGrowth ops · Marketplaces✓ readShip a ChatGPT app as a distribution channel
Build an MCP-based app for ChatGPT so the model can suggest your product when a conversation matches its tools.
WorkingSecondaryno examples0 examples · 4 sourcesGrowth ops · Process✓ readSolve the problem manually before writing code
Deliver the outcome by hand (reports, consulting, spreadsheets) for early customers until you understand the problem well enough to automate it.
WorkingSecondary1 example · 2 sourcesGrowth ops · Process✓ readTarget 100 happy customers
Set the early goal as a number of happy customers and test hypotheses about who they are by direct outreach, iterating fast.
WorkingPrimary20181 example · 2 sourcesGrowth ops · Billing✓ readUK subscription-contract regime
Prepare for UK rules on subscription cancellation and information due in January 2027.
RestrictedSecondaryno examples0 examples · 3 sourcesGrowth ops · Outbound✓ readVerified lists and steady send volume
Keep bounce rates under 2% by verifying contacts and send at a consistent pace to protect domain health.
WorkingVendor claim20251 example · 1 sourceGrowth ops · Outbound✓ readWaterfall enrichment across data providers
Query data vendors in cost order and stop at the first hit.
WorkingVendor claim1 example · 1 sourceGrowth ops · Process✓ readWeekly growth meeting with a growth lead
Hold a fixed weekly growth meeting run by a named growth lead, with a product or project owner on each test.
WorkingPrimaryno examples0 examples · 2 sourcesGrowth ops · Marketplaces✓ readWrite tool names and descriptions as the discovery surface
Treat MCP tool names, descriptions and example prompts as the main way assistants decide when to use you.
WorkingSecondaryno examples0 examples · 3 sources