TL;DR
- Shared vocabulary is infrastructure — treat it like one. Audit it, document it, onboard against it.
- Run it as a game — auction format makes a dry topic energizing and surfaces hidden gaps fast.
- Reward context, not spelling — anyone can recite DORA. Few can say what their team's MTTR is.
- End with one rule — a glossary, a "no jargon in onboarding week," anything observable.
A few sprints ago, I sat next to a new developer during refinement. He nodded through every story. Afterwards, in the kitchen, he confessed: "I thought WSJF meant 'We Should Just Finish.' I've been guessing for two weeks."
He's not the only one. Every team I've coached develops its own pidgin — a swirl of DORA, MTTR, WIP, OKR, mixed with BRB, SMH, ROFL — and assumes everyone on the call shares the same dictionary. They don't. Newer engineers fake it. Stakeholders smile and ask a colleague later. Shorthand is power, but it's also a gate.
Most retros optimize for output (what to keep / stop / start). Once a quarter, optimize for input instead — the shared vocabulary your team uses to think together. The cheapest alignment fix nobody runs.
What follows is the exact format I run when a team starts hiding behind acronyms. It takes 45–60 minutes, plays like a game, and ends with one or two concrete agreements about how the team will speak to each other from now on.
Retrospective Exercise: "Decode & Deliver"
A lighthearted, fast-paced retro format that helps the team master everyday cultural acronyms alongside critical DevOps and Agile metrics. It highlights how clarity in shorthand prevents communication bottlenecks.
Phase 1 — Slang Speed-Dating (10 min)
Goal: raise the team's energy and introduce the concept of shorthand.
- Set the stage: divide the team into pairs (breakout rooms if remote).
- The task: give the pairs 3 minutes to write a short, fictional Slack message to each other using only informal, everyday acronyms (e.g., "BRB, SMH looking at this PR, IMO it's a mess, TBH, LMK…").
- The reveal: bring everyone back. One person from each pair reads their cryptic message out loud while the rest of the team decodes it in real time in chat.
Phase 2 — The Acronym Auction (20 min)
Goal: learn and clarify agile, DevOps, and cultural terms in a competitive, playful setup.
- Preparation: before the session, split your board into two categories: "Watercooler Slang" and "The Engine Room" (Agile / DevOps). Hide the definitions.
- The rules: divide the team into two competing squads. The facilitator displays an acronym (e.g., WSJF or ROFL). The first team to buzz in (or type
MINE in chat) gets the chance to define it.
- Scoring:
- +1 point — correctly stating what the letters stand for.
- +2 points — correctly explaining how or when it should be used in the team's context.
- −1 point — if a team buzzes in but gets it wrong, the points go to the opposing team.
- Facilitator role: keep the pace brisk. When a technical acronym like DORA or MTTR comes up, take 30 seconds to reinforce why it matters to your specific pipeline before moving on.
Why context beats spelling
The +2 bonus is the whole point. Anyone can memorize that DORA stands for "DevOps Research and Assessment." Few can say what their team's current Change Failure Rate actually is, or whether MTTR was measured this quarter. The game rewards operational fluency, not trivia.
Phase 3 — Signal vs. Noise (15 min)
Goal: transition from a game into a traditional retrospective reflection on team communication.
Create a 2×2 matrix on the board with the following prompts. Give the team 5 minutes to add silent sticky notes, then a 10-minute open discussion.
Enablers
Which acronyms — technical or casual — help us move faster and communicate better?
Bottlenecks
Which acronyms cause confusion, exclusion, or alignment issues for new joiners or external stakeholders?
Overused / Misused
Are we weaponizing terms — overusing WIP limits without fixing flow, or misaligning on OKRs?
Team Agreement
What rules do we want to set for our shared vocabulary moving forward?
Phase 4 — Wrap-Up & Action Items (5 min)
Co-create one or two concrete action items. Keep them small, observable, and owned by a name.
Example action item: "Add an 'Acronym Glossary' section to our team's onboarding wiki page by Friday — owner: Maria."
The Reference Grid
A scannable cheat-sheet for the Auction phase — and for the kitchen-table moments when someone asks "wait, what does that mean?". Six grids by territory: chat slang, agile/product, devops/engineering, security/web, data/AI, and finance/banking.
Slang & Cultural
LOLLaughing Out Loud
ROFLRolling On the Floor Laughing
LMAOLaughing My A** Off
BRBBe Right Back
BTWBy The Way
TBHTo Be Honest
IDKI Don't Know
FYIFor Your Information
IMOIn My Opinion
IMHOIn My Humble Opinion
NGLNot Gonna Lie
FRFor Real
SMHShaking My Head
GTGGot To Go
IDCI Don't Care
TBFTo Be Fair
JKJust Kidding
TL;DRToo Long; Didn't Read
IRLIn Real Life
DIYDo It Yourself
BFFBest Friends Forever
TIAThanks In Advance
AFKAway From Keyboard
ASAPAs Soon As Possible
ETAEstimated Time of Arrival
TBDTo Be Determined
YOLOYou Only Live Once
HMUHit Me Up
DMDirect Message
FWIWFor What It's Worth
OOTDOutfit Of The Day
RTRetweet
AFAIKAs Far As I Know
RNRight Now
BCBecause
OMWOn My Way
Agile & Product
MVPMinimum Viable Product
MMFMinimum Marketable Feature
WIPWork In Progress
OKRObjectives and Key Results
KPIKey Performance Indicator
KRKey Result
WSJFWeighted Shortest Job First
MoSCoWMust / Should / Could / Won't (priority)
INVESTIndependent, Negotiable, Valuable, Estimable, Small, Testable
DoDDefinition of Done
DoRDefinition of Ready
PIProgram Increment
POProduct Owner
SMScrum Master
RTERelease Train Engineer
ICEImpact, Confidence, Ease (scoring)
RICEReach, Impact, Confidence, Effort
SAFeScaled Agile Framework
LeSSLarge-Scale Scrum
XPExtreme Programming
BAUBusiness As Usual
RACIResponsible, Accountable, Consulted, Informed
NPSNet Promoter Score
DAUDaily Active Users
DevOps & Engineering
CI/CDContinuous Integration / Continuous Deployment
DORADevOps Research and Assessment (4 metrics)
MTTRMean Time To Resolution
MTBFMean Time Between Failures
SRESite Reliability Engineering
IaCInfrastructure as Code
K8sKubernetes
PR / MRPull Request / Merge Request
LGTMLooks Good To Me (review)
TDDTest-Driven Development
BDDBehavior-Driven Development
DDDDomain-Driven Design
YAGNIYou Ain't Gonna Need It
DRYDon't Repeat Yourself
KISSKeep It Simple, Stupid
SOLIDSingle, Open-closed, Liskov, Interface, Dependency
RESTRepresentational State Transfer
CRUDCreate, Read, Update, Delete
ACIDAtomicity, Consistency, Isolation, Durability
ORMObject-Relational Mapping
APIApplication Programming Interface
SDKSoftware Development Kit
IDEIntegrated Development Environment
SSOSingle Sign-On
JWTJSON Web Token
OAuthOpen Authorization
SLAService Level Agreement
SLOService Level Objective
Security & Web
CSRFCross-Site Request Forgery
XSSCross-Site Scripting
CORSCross-Origin Resource Sharing
SQLiSQL Injection
RBACRole-Based Access Control
ABACAttribute-Based Access Control
MFAMulti-Factor Authentication
2FATwo-Factor Authentication
SAMLSecurity Assertion Markup Language
OIDCOpenID Connect
TLSTransport Layer Security
SSLSecure Sockets Layer
VPNVirtual Private Network
IAMIdentity and Access Management
PIIPersonally Identifiable Information
GDPRGeneral Data Protection Regulation
OWASPOpen Worldwide Application Security Project
DoSDenial of Service
DDoSDistributed Denial of Service
ZTAZero Trust Architecture
PKIPublic Key Infrastructure
MITMMan-in-the-Middle (attack)
CVECommon Vulnerabilities and Exposures
CVSSCommon Vulnerability Scoring System
WAFWeb Application Firewall
IDSIntrusion Detection System
IPSIntrusion Prevention System
SIEMSecurity Information and Event Management
SOCSecurity Operations Center
EDREndpoint Detection and Response
XDRExtended Detection and Response
SBOMSoftware Bill of Materials
Data & AI
AIArtificial Intelligence
MLMachine Learning
LLMLarge Language Model
RAGRetrieval-Augmented Generation
NLPNatural Language Processing
CVComputer Vision
DNNDeep Neural Network
CNNConvolutional Neural Network
RNNRecurrent Neural Network
GPTGenerative Pre-trained Transformer
RLHFReinforcement Learning from Human Feedback
AGIArtificial General Intelligence
MCPModel Context Protocol
MLOpsMachine Learning Operations
ETLExtract, Transform, Load
ELTExtract, Load, Transform
OLAPOnline Analytical Processing
OLTPOnline Transaction Processing
SQLStructured Query Language
NoSQLNot Only SQL
BIBusiness Intelligence
EDAExploratory Data Analysis
DWHData Warehouse
TPUTensor Processing Unit
GPUGraphics Processing Unit
CUDACompute Unified Device Architecture
ANNApproximate Nearest Neighbor (vector search)
Finance & Banking
KYCKnow Your Customer
KYBKnow Your Business
AMLAnti-Money Laundering
CFTCombating the Financing of Terrorism
CDDCustomer Due Diligence
EDDEnhanced Due Diligence
CIPCustomer Identification Program
SARSuspicious Activity Report
PEPPolitically Exposed Person
OFACOffice of Foreign Assets Control (US sanctions)
PCI DSSPayment Card Industry Data Security Standard
SWIFTSociety for Worldwide Interbank Financial Telecom
IBANInternational Bank Account Number
BICBank Identifier Code
ACHAutomated Clearing House
SEPASingle Euro Payments Area
POSPoint of Sale
ATMAutomated Teller Machine
APRAnnual Percentage Rate
APYAnnual Percentage Yield
SOFRSecured Overnight Financing Rate
FXForeign Exchange
BNPLBuy Now, Pay Later
AUMAssets Under Management
IPOInitial Public Offering
ROIReturn on Investment
MRRMonthly Recurring Revenue
ARRAnnual Recurring Revenue
LTVLifetime Value
CACCustomer Acquisition Cost
If you try this with your team, adjust the technical list to match your stack. Replace DORA with SPACE if that's your model. Add your platform's house terms — every team has half a dozen acronyms that only exist inside the building. Those are the ones most worth decoding.