The four layers: vision, mission, strategy, and tactics

Description

You already have two of the four pieces: vision (where we're headed, with no deadline) and mission (why we exist today). This lesson completes the map by adding the two layers that follow, which you'll use throughout the rest of this guide: strategy (the set of choices about where to play and how to win, which you already saw in module 1) and tactics (what gets built this week, this iteration, with a feature name). The four stack in a strict order — vision → mission → strategy → tactic — and the property connecting them is the one that names this lesson: each layer constrains what makes sense in the layer below it. You can't choose just any strategy and expect it to serve just any vision; you can't build just any tactic and expect it to serve just any strategy.

Connection to the module. Lessons 2 and 3 gave you the definitions, one at a time, of the two top layers. This lesson doesn't add new content about vision or mission — it brings them together with strategy and tactics into a single map, so you can place any statement you hear in a meeting ("we're going to focus on X," "this sprint we're launching Y") in its correct layer, without confusing it with the one next door. Lesson 5 uses this map to explain why a useful vision also rules things out — but before ruling anything out, you need to be able to name which layer everything lives in. That's what this lesson gives you.

An everyday analogy: the map's zoom levels

Open any map app and zoom out until you see an entire continent. At that level you see the world's general shape — where the oceans are, where the mountain ranges are — but you see no streets, not even any legibly named cities. That's the vision level: the general shape of where everything is headed, with no operational detail. Now zoom in a bit: countries and regions appear — you still can't see streets, but you now know what territory you're in. That's the mission level: why you're in this territory and not another, today.

Zoom in again and the city appears, with its neighborhoods and main avenues — you can trace a general route across it. That's the strategy level: the choices of where to play (which neighborhoods) and how to win (which avenues to take) within the territory you already chose. And if you zoom in once more, you see the exact street, the traffic light, the next right turn. That's the tactics level: the concrete step happening right now.

The key property of a zoomable map is the one that names this lesson: you can't pick a street that isn't in the city you chose, or a city that isn't in the country you're headed toward. Each zoom level constrains the valid options at the next, closer level. A right turn, taken in isolation, is neither good nor bad — it only makes sense, or doesn't, depending on whether it brings you closer to the city you chose, which in turn only makes sense depending on whether it's in the direction of your continent. That is exactly what happens to an engineering tactic evaluated outside its strategy, or a strategy evaluated outside its vision: without the layers above, there's no way to know if it's a good turn.

Worked example: Mercado's 4 layers, classified

We gather, in one place, a real statement from each of Mercado's four layers, and use classifyLayer — the model you've been using since lesson 1 — to verify that each one falls where it should.

// Pedagogical model: recognizes linguistic signals typical of each layer.
// It doesn't understand the statement -- it recognizes the vocabulary each
// layer tends to be expressed in (choice verb, action verb, "today",
// horizon with no date).
function classifyLayer(statement) {
  const text = statement.toLowerCase();
  const signals = {
    tactic: ['launch', 'implement', 'add', 'sprint', 'this quarter', ' q1', ' q2', ' q3', ' q4'],
    strategy: ['we choose', 'instead of', 'we focus on', 'we will invest', 'the next', 'we bet on', 'we win'],
    mission: ['we exist to', 'our mission', 'we help', ' today'],
    vision: ['the world where', 'the place where', 'someday', 'our vision', 'we imagine a'],
  };
  const order = ['tactic', 'strategy', 'mission', 'vision'];
  for (const layer of order) {
    if (signals[layer].some((kw) => text.includes(kw))) {
      return layer;
    }
  }
  return 'unclear';
}

const statements = [
  { label: 'Vision', text: 'The world where anyone discovers, on Mercado, what they didn\'t know they wanted.' },
  { label: 'Mission', text: 'We exist to connect curious buyers with trusted local sellers, today.' },
  { label: 'Strategy', text: 'We choose to focus on buyers who browse, not those searching for an exact SKU, and we win by investing in curated discovery instead of the lowest price.' },
  { label: 'Tactic', text: 'This quarter we launch the redesigned recommendations feed in checkout.' },
];

console.log('=== Mercado\'s 4 layers, classified ===\n');
for (const s of statements) {
  const layer = classifyLayer(s.text);
  console.log(`[${s.label}] "${s.text}"`);
  console.log(`  -> classifyLayer(): ${layer}\n`);
}

What to expect. Running the file with Node, the output is exactly this:

=== Mercado's 4 layers, classified ===

[Vision] "The world where anyone discovers, on Mercado, what they didn't know they wanted."
  -> classifyLayer(): vision

[Mission] "We exist to connect curious buyers with trusted local sellers, today."
  -> classifyLayer(): mission

[Strategy] "We choose to focus on buyers who browse, not those searching for an exact SKU, and we win by investing in curated discovery instead of the lowest price."
  -> classifyLayer(): strategy

[Tactic] "This quarter we launch the redesigned recommendations feed in checkout."
  -> classifyLayer(): tactic

The four outputs confirm the complete map. Now look at it as a single chain, top to bottom, with the "constrains" relationship made explicit at each arrow:

VISION      "The world where you discover what you didn't know you wanted"
  |          horizon: 10-25 years -- no deadline -- almost never changes
  v  constrains which missions make sense
MISSION     "We exist to connect buyers and local sellers, today"
  |          horizon: today, with what already exists -- very stable
  v  constrains which strategies make sense
STRATEGY    "We choose buyers who browse, not exact-SKU search;
             we win with curated discovery, not the lowest price"
  |          horizon: 1-3 years -- changes if the competitive game changes
  v  constrains which tactics make sense
TACTIC      "This quarter we launch the redesigned recommendations
             feed in checkout"
             horizon: days/weeks -- changes every sprint or quarter

Read the full chain as a cascading justification: the recommendations feed redesign (tactic) makes sense because it serves the strategy of winning with curated discovery; that strategy makes sense because it serves the mission of connecting curious buyers with sellers today; and that mission matters because it points toward the vision of a world where shopping starts with discovering. Remove any link and the layers below it float with no anchor — a tactic nobody can connect upward is, literally, work with no strategic justification, no matter how well it's executed.

Going deeper: the rate of change isn't the same across the 4 layers

An almost invisible mistake is treating the four layers as if they changed at the same pace. They don't, and the difference in speed is, itself, a useful signal for knowing which layer you're standing on:

• Vision: changes almost never. If your company changes its vision every year, you didn't have a vision — you had a strategy with a pompous name (the exact mistake from lesson 2).
• Mission: changes very little, almost as stable as the vision, because it describes something that's already true today.
• Strategy: changes with the game. New competitors, new technology, a market that saturates — any of these can justify a new strategy, without the vision or mission moving an inch. This is healthy, not a sign of indecision.
• Tactics: changes all the time — every sprint, every experiment result, every backlog learning. It's the most disposable of the four layers, and that's how it should be: if your tactics never change, you're probably not learning anything from the market.

A quick way to self-diagnose which layer you're standing on in a conversation: ask yourself "will we still be saying this exact thing a year from now?". If the answer is "definitely yes," you're probably in vision or mission. If it's "depends how this quarter goes," you're in strategy. If it's "I probably won't even remember this in a month," you're in tactics — and that's perfectly fine, as long as you can trace it, upward, to the strategy it serves.

Common mistakes

Confusing the layer by tone, not by content. What happens: a well-written tactic, with inspiring language ("we're going to revolutionize checkout"), slips into a conversation as if it were strategy or even vision, just because it sounds big. Why it happens: aspirational tone gets associated, for no good reason, with the layers above — but tone doesn't determine the layer, the horizon and the subject do. How to spot it: ask "would this still be true in a year, no matter what happens in the market?" If the answer depends on a specific launch this quarter, it's a tactic, no matter how grandiose it sounds. How to fix it: use classifyLayer or lesson 2's three-question test — aspirational vocabulary isn't a reliable signal; the time horizon is.

Designing a strategy that doesn't connect upward to any vision or mission. What happens: a team puts together an ambitious strategic plan — new markets, new business lines — without checking whether that plan serves the existing vision, or actually contradicts it. Why it happens: strategy often gets designed in response to competitive pressure or a short-term opportunity, without going back to look at the vision document that's been filed away for a year. How to spot it: for any new strategic choice, try completing the sentence "this serves our vision because...". If you can't finish it honestly, the strategy is floating, disconnected from the layer that gives it meaning. How to fix it: before approving a new strategy, explicitly trace it up to the vision and mission — the same tracing exercise you did in product-thinking-for-engineers-guide to connect a task to the business, applied one level up.

Treating strategy as if it had the same immutability as vision. What happens: a team clings to a three-year-old strategy as if changing it meant betraying the company's vision — "we've always competed this way, we can't change now." Why it happens: the correct stability of vision (which almost never changes) gets unconsciously mixed up with an incorrect stability attributed to strategy (which should change when the competitive game changes). How to spot it: if defending an old strategy starts sounding like defending the vision — "this is who we are" — you're probably confusing layers. How to fix it: remember this lesson's speed table — vision is the one that almost never changes; strategy changes with the market, and changing it isn't disloyalty to the vision, it's exactly what strategy is there to do.

Exercises

Exercise 1 — Classify without running the code. For each statement, predict what classifyLayer would return:

• (a) "We will invest the next two years in deepening our relationship with local sellers, instead of competing on price with the big players."
• (b) "Add a featured-sellers banner to the homepage."
• (c) "Our mission is to make every purchase start with curiosity, not a list."

See solution

• (a) strategy. Contains "we will invest" and "instead of," both strategy-layer signals — it describes a choice of where to focus resources, with a horizon of a couple of years.
• (b) tactic. Contains "add," a concrete action with a feature name, with no mention of why or which strategy it points toward — typical of an isolated tactic.
• (c) mission. Contains "our mission," the most direct signal of the layer. Notice that, even though the content sounds almost like vision (it talks about "every purchase"), the explicit linguistic signal ("our mission") is what decides — another reason not to rely on tone alone, as you saw in common mistakes.

Exercise 2 — Trace the chain upward. For the tactic "this quarter we launch the redesigned recommendations feed in checkout," write, in one sentence each, the Mercado strategy and mission that justify it — without inventing anything new, using what you already saw in this lesson.

See solution

The strategy that justifies it: "We choose to focus on buyers who browse, not those searching for an exact SKU, and we win by investing in curated discovery instead of the lowest price." (The checkout recommendations feed is, literally, curated discovery in action.)

The mission that justifies it, one level up: "We exist to connect curious buyers with trusted local sellers, today." (A well-curated recommendations feed helps that connection right now, without needing to wait for the 20-year vision.)

The full chain, bottom to top: the tactic serves the strategy, which serves the mission, which points toward the vision. If any of those links broke — if the redesign had nothing to do with "curated discovery" — the tactic would be left without justification, no matter how perfectly it was executed.

Exercise 3 — Spot the wrong layer. A Mercado engineer defends a task by saying: "This is part of our vision of being the technically fastest marketplace." Which layer is the word "vision" being misused for, and what would be the correct name for what's actually being described?

See solution

They're using "vision" for what is actually, at most, a technical strategy (or even an architecture preference) — "be the technically fastest" describes a choice about how to compete, not a state of the world with a 10-25-year, dateless horizon, nor something a buyer directly experiences as "the world changed." Applying lesson 2's test: does it survive a strategy change? If Mercado decided to compete on curation instead of speed, "be the fastest" would stop making sense — that confirms it's strategy (or infrastructure tactics), not vision. The correct name, in the right vocabulary, would be something like: "this task serves our strategy of reducing friction in discovery" — much more precise, and much easier to honestly drop if it stops serving.

Summary and next step

The four layers — vision, mission, strategy, tactics — stack in a strict order, each one constraining what makes sense in the one below it, and each one changing at a different pace: vision almost never, mission rarely, strategy with the market, tactics every sprint. You saw classifyLayer place a real statement from each of Mercado's layers, and you saw the full chain traced top to bottom, with each arrow explaining "why this justifies what's below it."

Before moving on you should be able to: name the four layers in order, with their approximate time horizon; and for any statement ("we're going to...," "we're going to make...," "we exist to..."), place it in its correct layer without being fooled by tone.

Lesson 5 uses this complete map to answer the question that gives the whole module its name: if a real vision constrains everything beneath it, what does a vision that does that job look like — a vision that says no to concrete things, not just yes to everything nice-sounding?

Resources

• Roger Martin, "Decoding the Strategy Choice Cascade" — rogermartin.medium.com/decoding-the-strategy-choice-cascade-475d40555eb1. Playing to Win's five cascading choices are, in essence, a finer breakdown of the "each layer constrains the one below" relationship you saw in this lesson. In English.
• Richard Rumelt, Good Strategy Bad Strategy — goodbadstrategy.com. Strategy's "kernel" (diagnosis, guiding policy, coherent actions) is, in this module's vocabulary, a well-built strategy layer — with actions (tactics) that genuinely connect to the guiding policy. In English.
• Marty Cagan (SVPG), "Vision vs. Strategy" — svpg.com/vision-vs-strategy. The same layer separation, applied specifically to software product teams. In English.
• Gibson Biddle, "The DHM Model" — gibsonbiddle.medium.com/2-the-dhm-model-6ea5dfd80792. How Netflix's former VP of Product connects a product vision to concrete strategies, without skipping layers. In English.