Library
00/07 · ~34 min
GUIDEDECK · how teams actually ship

Agile & Delivery
— ship small, learn fast.

A 34-minute working session on the way modern teams turn ideas into shipped software — the agile mindset, Scrum and Kanban, estimation, writing tickets people can actually build, and the tools that hold it together.

~34 MINBEGINNER → INTERMEDIATEPROCESS-AGNOSTIC
SCROLL
01 · Why agile 4 min

Ship small and learn fast,
instead of one big-bang release.

The hard part of software isn't typing — it's building the right thing. You rarely know exactly what users need up front, so the safest plan is to deliver a little, get real feedback, and steer. Agile is just that loop, made into a habit.

Agile a mindset for delivering in small steps— means breaking work into thin slices that each produce something usable, showing it to people, and adjusting from what you learn. It's the opposite of locking a year-long plan in stone and hoping it still fits reality at the end.
WATERFALL · value once, at the end Plan Build Test Ship first feedback here AGILE · value every increment Slice 1build·ship·learn Slice 2build·ship·learn Slice 3build·ship·learn feedback steers the next slice

Waterfall bets everything on one release; agile ships thin slices and lets feedback redirect the next one.

The manifesto, in plain words

In 2001 a group of practitioners wrote the Agile Manifesto. Stripped of buzzwords, it values:

  • People over process — talking to each other beats following a rigid procedure.
  • Working software over documents — a running demo proves more than a spec.
  • Collaboration over contracts — work withthe customer, don't negotiate against them.
  • Responding to change over a plan — when reality shifts, update the plan instead of defending it.

Like cooking a new dish — you taste as you go and adjust the seasoning, rather than serving an untasted banquet.

Heads up — agile is a mindset, not a tool you install. Scrum and Kanban (next) are the two most common ways to practiceit. You can run either one badly; the values above are what tell you if it's actually working.
02 · Scrum 6 min

Work in fixed sprints with
clear roles and a steady rhythm.

Scrum is the most popular agile framework. It packages work into short, fixed time-boxes and adds a small set of roles and meetings so the team inspects and adapts on a predictable beat.

Scrum an agile framework built on sprints — a team commits to a small batch of work, delivers it in a fixed window (the sprint), then reviews and adjusts before the next one. It defines three roles, a handful of events, and three artifacts — nothing more.
Sprint a fixed time-box, usually 1–2 weeks, in which the team builds a usable increment. The length never changes, so the team gets a reliable heartbeat and a regular point to course-correct.
Product Backlog Sprint Planning SPRINT 1–2 weeks daily scrum ↻ Incre- ment Review Retro learnings feed the backlog

One sprint: pull from the backlog, plan, build with a daily check-in, ship an increment, then review and retro — and repeat.

The five events (ceremonies)

  • Sprint Planning — pick what the team will build this sprint and agree on a goal.
  • Daily Scrum (standup)— a 15-minute sync: what moved, what's next, what's blocked. It is not a status report to a boss.
  • Sprint Review — demo the working increment to stakeholders and gather feedback.
  • Retrospective — the team reflects on how it worked and picks one or two improvements.
  • The sprint itself is the container that holds the other four.

Like a TV season: a fixed run of episodes, a writers' room each morning, a finale demo, then notes before the next season.

The three roles

Product Owner

Owns the what & the why

Decides priorities and keeps the backlog ordered by value. One person, the single voice on "what matters most next" — so the team never gets five conflicting roadmaps.

Scrum Master

Owns the how it flows

A coach and blocker-remover, not a manager. Protects the team from distractions, facilitates the events, and helps the process improve.

Developers

Own the how it's built

The cross-functional people who design, build, and test the increment. They decide how much fits in a sprint and how to deliver it.

Velocity how much work a team finishes per sprint, measured in story points (Part 4). It's a planning aid for the team's own forecasting — not a productivity score to compare teams or push people harder.
03 · Kanban 5 min

Continuous flow — limit
work in progress, finish things.

Where Scrum batches work into sprints, Kanban drops the time-boxes and optimizes for steady flow: visualize the work, cap how much is in progress at once, and pull the next item only when there's room.

Kanban a flow-based way to manage work— you put every item on a board, move cards left to right as they progress, and limit how many can sit in each stage. There are no fixed sprints; work flows continuously and you ship whenever something's done.
WIP limit a cap on items in progress at once(e.g. "max 3 In Progress"). When a column is full, you can't start anything new — you must help finish what's already there. Less juggling, faster delivery.
To Do In Progress · 3 Review · 2 Done blocked: full pull right only when the next column has room

Cards flow left → right. "In Progress" is capped at 3 — a fourth card can't start until one moves on, so the team finishes instead of starting.

items time → Done In Progress To Do

A cumulative flow diagram stacks the bands over time — a widening In Progress band warns you work is piling up faster than it finishes.

Reading flow, two numbers

  • Cycle time— how long one item takes from "started" to "done". Lower is faster delivery.
  • Throughput — how many items finish per week. This is what you forecast with.
  • A bottleneckshows up as a column that's always full — the WIP limit makes it visible so you can swarm it.
  • Kanban shines for steady, unplanned work — support, ops, a stream of small fixes — where fixed sprints feel forced.
04 · Backlog & estimation 5 min

A prioritized list of user stories,
sized by effort — not by hours.

The backlog is the team's single to-do list, ordered so the most valuable work sits on top. Each item is written from the user's point of view, and the team sizes the big ones so it can forecast.

Backlog the ordered list of everything to do: features, fixes, and improvements. It's living — re-ordered as priorities shift. Only the top items need to be detailed; the rest can stay rough until they rise ("backlog refinement").
User story a feature told from the user's view, in one line: As a [user], I want [goal], so that [benefit]. It captures the why, not the technical solution — leaving the team room to decide how.
// the template As a shopper I want to save items to a wishlist so that I can buy them later // good stories are vertical slices — // a thin bit of UI + logic + data that // delivers value on its own.
Epic Story Story Story task task

A big epic splits into stories (user-facing slices), which split into tasks (the dev steps).

Story points a relative size for effort, not a count of hours. They fold together complexity, uncertainty, and volume into one number. Teams use them because humans are far better at saying "this is twice as big as that" than "this is 14.5 hours".
relative size · Fibonacci-ish 1 2 3 5 13 too big — split it

Gaps widen on purpose: big items are fuzzy, so you don't pretend to tell 8 from 9. A 13 is a signal to split, not to estimate harder.

Why points beat hours

  • Hours pretend to be preciseand they vary by person — a senior's 2 hours is a junior's morning.
  • Points compare, and the team's velocity turns them into a forecast over a few sprints — no per-person guessing.
  • Estimate together. Planning poker surfaces hidden complexity: when two people pick 2 and 8, the gap is the conversation.
  • Points are a team forecasting tool, never a stick to beat people with.

Prioritization — what to do first

MoSCoW

Must / Should / Could / Won't

Sort items into four buckets so "everything is top priority" stops being an option. The Won't bucket is the honest one.

Value vs effort

Chase the cheap wins

Plot value against effort; do high-value, low-effort first. It turns gut feeling into a picture the whole team can argue with.

WSJF

Weighted shortest job first

Roughly: value ÷ effort. Favors small, urgent, high-value work — a tie-breaker when the backlog is long.

05 · Writing good tickets 5 min

Small, clear, testable —
with acceptance criteria.

A ticket is a promise about a small piece of work. A good one tells anyone on the team what to build and how you'll know it's done — without a meeting to decode it.

Acceptance criteria the conditions that make a ticket "done". Written as plain, checkable statements (often Given / When / Then), they remove ambiguity: the developer knows when to stop, and the reviewer knows what to verify.
Vague — a meeting waiting to happen
Title: Fix the login Description: Login is broken sometimes, please look into it and make it better. Thanks! // who? which login? what's "better"? // no way to tell when this is done.
Clear — buildable & checkable
Title: Show error on wrong password As a returning user I want a clear message on a bad login so that I know to retry, not that the site broke Acceptance criteria: Given a registered email + wrong password When I submit the form Then I see "Incorrect email or password" and the password field is cleared
IIndependent — stands on its own NNegotiable — the how is open VValuable — a user gains something EEstimable — the team can size it SSmall — fits in one sprint TTestable — clear pass/fail

INVEST — a quick checklist for whether a story is ready to pull into a sprint.

Habits that save hours

  • One ticket, one outcome.If the title needs an "and", it's probably two tickets.
  • Write the criteria first.If you can't describe "done", the work isn't understood yet.
  • Add context, link the source. The bug report, the design, the Slack thread — future-you will thank present-you.
  • Agree a Definition of Donefor the whole team (tests pass, reviewed, deployed) so "done" means the same thing every time.

Like a recipe card: ingredients, steps, and a photo of the finished plate — not just "make dinner".

06 · The tooling 5 min

Where the work and the
knowledge actually live.

Two categories cover most teams: an issue tracker for the work items and a knowledge basefor the docs. The tool matters far less than using it consistently — but here's the honest landscape.

Where tickets, boards & sprints live

Jira

The enterprise standard

Pro — endlessly configurable; handles big orgs, complex workflows, and reporting.

Con — that power gets heavy and slow; easy to over-configure into a maze.

Linear

Fast & opinionated

Pro — keyboard-driven, quick, opinionated defaults that keep teams moving.

Con — fewer escape hatches for unusual processes; less mature for huge enterprises.

GitHub Projects

Lives with the code

Pro — free with GitHub; issues sit right next to PRs and commits.

Con — lighter on agile reporting and cross-team rollups than dedicated tools.

Where decisions & how-tos live

Confluence

The Atlassian wiki

Pro — structured spaces, strong permissions, and tight Jira integration for big orgs.

Con — editing feels dated and clunky next to newer tools; search can frustrate.

Notion

Flexible docs + databases

Pro — fast, friendly editing; docs and lightweight databases in one place.

Con — that flexibility sprawls without discipline; weaker for strict enterprise governance.

Match the tool to the team, not the hype

  • Already on GitHub, small team? Start with GitHub Projects — zero new tools, code and work in one place.
  • Product team that values speed? Linear — minimal setup, fast by default.
  • Large org, many teams, audit & reporting needs? Jira + Confluence — the integration and governance pay off at scale.
  • Docs: reach for Notion when you want frictionless writing, Confluence when you need structure and permissions.
  • Whatever you pick: one source of truth, kept current, beats the "best" tool that's half-abandoned.
Don't let the tool run the team. The board should mirror how you actually work, not force a 12-step workflow nobody follows. If updating the tracker feels like a chore that adds no clarity, simplify it — the ceremony serves the team, never the reverse.
07 · Scrum vs Kanban — choosing & recap 4 min

Pick the rhythm that fits
the work in front of you.

They're not rivals — they're different tools. Scrum suits planned, batchable work with stakeholders who want a cadence; Kanban suits a steady stream of varied, unplanned work. Many teams blend them.

Lean toward Scrum when…
  • Work can be planned in batches a sprint ahead.
  • Stakeholders want a predictable demo cadence.
  • The team benefits from a regular rhythm and reset.
  • You're building toward features and roadmaps.
Lean toward Kanban when…
  • Work arrives unpredictably (support, ops, bugs).
  • Priorities change faster than a sprint.
  • You want to optimize flow and cycle time.
  • Fixed time-boxes feel like overhead, not help.
Scrumban a common hybrid— keep Scrum's planning and retro, but add Kanban's board and WIP limits and drop the strict commitment. Most real teams land somewhere on this spectrum rather than running either by the book.

Five things to walk out with

1Ship small, learn fast. Thin slices plus real feedback beat a big-bang plan every time.
2Scrum = sprints + roles + ceremonies; Kanban = continuous flow with WIP limits. Both are agile.
3Size with points, not hours. Velocity is a team forecast, not a productivity score.
4A good ticket is small, clear, and testable— with acceptance criteria that define "done".
5The process serves the team. Pick the lightest one that keeps you shipping — and keep improving it in the retro.
Knowledge check

Did it stick?

Five quick questions on agile, Scrum, Kanban, estimation, and tickets — instant feedback, no sign-in.

Rate this deck
be the first

Navigate with ← → or scroll · back to library