How the leader track works
Six sessions, zero spreadsheets. You learn the GMV equation and who owns each lever today (a1), get a working dictionary of the ten metrics your team throws around (a2), learn to read a funnel chart in under a minute (a3), run an Amazon-style weekly business review (a4), judge your numbers against honest industry benchmarks and seasonality (a5), and finish with metric governance - the rules that stop two teams reporting two different "conversion rates" to the same board (a6). The analyst track (b1-b8) computes everything by hand on the same store; this track teaches you what to ask, not what to calculate.
GMV = Traffic x Conversion x AOV 12 min live
This course runs on one fictional store, Mango Lane - a mobile-first fashion app - so every number holds still and cross-checks. In a typical month it gets 360,000 sessions (visits), 2.7% of those sessions end in an order, and the average order is worth $68. Multiply the three and you get the month's top line: $660,960 of GMV (gross merchandise value - the total value of goods sold, before anyone argues about returns and discounts; a2 handles the arguing). Three numbers in, one board number out. There is no fourth lever hiding anywhere.
LiveWhy leaders love this equation (and finance quietly does too)4 min▶
Multiplying a few operational drivers into one headline number is an old, respectable trick: DuPont was decomposing return on equity into margin x turnover x leverage a century ago, and every modern growth model is the same move applied to revenue. The power is not the arithmetic - it is that each factor lands on a different team's desk:
- Traffic (sessions) - how many visits happened. Bought, earned, or emailed into existence by marketing.
- Conversion rate - what share of visits ended in an order. Won or lost in the shopping experience, so it belongs to product and UX.
- AOV - what an average order was worth. Shaped by pricing, bundling, and assortment: merchandising.
When GMV moves, exactly one question matters first: which of the three moved? Everything else in this track builds on that reflex.
Self-studyThe same equation works at any altitude3 min read▶
Traffic x conversion x AOV is not just a whole-company view - it holds for any slice you point it at. Each of Mango Lane's four channels has its own three levers and its own mini-GMV:
- Organic - $264,384 (40% of GMV)
- Paid - $198,288 (30%)
- CRM (email & push) - $132,192 (20%)
- Social - $66,096 (10%)
The same is true per category, per country, per campaign. When a leader asks "why is paid underperforming?", the answer is always the same three-lever read, one altitude down. The analyst track's b6 (segment cuts) computes these slices; you just need to know they exist and sum to the total.
You cannot "do GMV" - you can only pull levers 10 min live
GMV is an outcome. No team can walk into the office and work on GMV directly, any more than a football team can work on "the score". Teams can only pull levers - run a campaign (traffic), fix a clunky checkout (conversion), bundle products (AOV) - and the outcome follows. A target set on the outcome without naming a lever is a wish. This distinction sounds obvious and is violated in roughly every second board deck.
LiveThe pass-through idea in plain words4 min▶
Because the three levers are multiplied, a change in any one of them passes straight through to the top line. If traffic falls 10% and nothing else changes, GMV falls about 10%. If AOV rises 10%, GMV rises about 10%. There is no cushion and no delay - the equation has no shock absorber.
Two useful consequences for a leader:
- Small lever moves are never small. A "minor" 0.3-point conversion dip at Mango Lane (2.7% to 2.4%) is about an 11% haircut on the lever - and therefore about an 11% haircut on GMV. Roughly $73,000 gone in a month, from a change that looks like rounding.
- Lever moves combine. Traffic up 10% while conversion is down 10% roughly cancels out - the top line looks flat while two big stories hide underneath. Flat GMV is not automatically good news; a3 teaches you to look underneath.
Self-studyWhy each lever needs an owner BEFORE it needs a target4 min read▶
The natural instinct is to set targets first: "conversion to 3% by Q3." But a target on an unowned lever produces a predictable failure mode - when it slips, the review meeting becomes an archaeology dig into whose fault it was, and every team has a plausible alibi (marketing blames site speed, product blames traffic quality, both are sometimes right).
Ownership first reverses the dynamic:
- An owner watches the lever weekly - so a dip is caught in days, not discovered at quarter end.
- An owner can explain a move - "conversion dipped because the new checkout shipped Tuesday" is only available from someone who owns the number and the roadmap behind it.
- Only then does a target mean anything - a target is a promise, and promises need a person attached.
Your homework today is exactly this: levers and owners, no targets yet. Targets come in a5 when you can see benchmarks.
Three moves, three phone calls ★ 15 min · everyone reads
No math here - just reading. Three things happen at Mango Lane. For each one: which lever moved, which team you call, and what you ask. This is the whole leadership skill: right lever, right owner, right question, in that order.
LiveScenario 1 - sessions dip 10% after an ad budget cut5 min▶
Finance trimmed paid spend mid-month. Two weeks later the weekly report shows sessions down about 10%, and GMV down about the same. Nobody should be surprised - a traffic lever move passes straight through - but somebody should own the response.
Which lever: traffic. Conversion and AOV are steady, so the top line fell purely because fewer visits happened.
Who you call: marketing - traffic is their lever, and the budget cut hit their channel mix.
What you ask: "Which channel lost the sessions, and were they sessions that used to convert?" Cheap traffic converting at 0.5% and CRM traffic converting at 5% are not the same loss - the answer decides whether the cut was smart or expensive.
LiveScenario 2 - conversion dips after a checkout redesign5 min▶
Product shipped a sleeker checkout on the 1st. By the 10th, conversion has drifted from 2.7% toward 2.4%. Traffic is normal, AOV is normal - GMV is quietly bleeding through one lever, and the timing is not a coincidence until proven otherwise.
Which lever: conversion. Same number of visits, same order size - a smaller share of visits is finishing the purchase.
Who you call: product and UX - the shopping experience is their lever, and they just changed it.
What you ask: "Where in the journey are shoppers now dropping, compared to before the release - and can we roll back while we investigate?" You do not need the funnel vocabulary yet (a3 gives it to you); you need the date of the change and a before/after picture.
LiveScenario 3 - AOV rises after bundling5 min▶
Merchandising launched "complete the look" bundles. AOV climbs from $68 toward $75, and GMV rises with it. Good news - but good news deserves the same discipline as bad news, because you want to know whether it will still be there next quarter.
Which lever: AOV. Same visits, same share buying - each order simply got bigger.
Who you call: merchandising - to congratulate them, and then to interrogate the win.
What you ask: "Is the lift from more items per order or pricier items - and did conversion hold while AOV rose?" Bundles sometimes lift order size while scaring off smaller buyers; a real win moves one lever without quietly denting another.
Self-studyWhen two levers move at once3 min read▶
Real months are rarely as clean as the three scenarios above. A sale event typically pushes traffic and conversion up while pulling AOV down; an ad cut pulls traffic down while conversion drifts up (the remaining visitors are the keen ones). Two habits keep multi-lever months readable:
- Ask for all three levers every time, even when the answer seems obvious. The lever nobody mentioned is where surprises hide.
- Ask which lever moved because of us and which moved on its own. A deliberate trade (discount lowers AOV, lifts conversion) is strategy; the same picture arriving uninvited is a problem.
a3 gives you the funnel view that makes multi-lever months quick to read, and a4 puts the reading on a weekly rhythm so no lever drifts unwatched for long.
The pattern to keep: in all three scenarios the first question was never "why is GMV down/up?" - it was "which lever moved?" GMV itself never gets a phone call. Levers have owners; outcomes only have meetings.
Try it yourself - this week ◐ 15-20 min total
- Write your own business's top line as the three levers: traffic x conversion x average order (or your industry's equivalents - visits, win rate, deal size work the same way). Rough numbers are fine; the structure is the exercise.
- Next to each lever, write one named owner - a person, not a team. Note where you had to hesitate: hesitation is the finding.
- Ask your data team the Part 1 question: last month's three factors next to the month before, on one slide.
- Bring your three-levers-with-owners sheet to a2 - the metric dictionary will sharpen every word on it.
Sources this session draws on
Ecommerce metrics have no single certificate - the definitions live in the analytics platforms' documentation and a century of decomposition practice. This page draws on:
Three questions before you go 🎯 ◐ 90 seconds
1 · Which of these is an outcome rather than a lever?
GMV is the result of the three levers multiplied - no team can work on it directly. Sessions, conversion rate, and AOV are the levers teams actually pull.
2 · Traffic falls 10% while conversion rate and AOV stay flat. What happens to GMV?
The levers are multiplied, so a 10% drop in one factor (with the others flat) is about a 10% drop in the product. No cushion, no amplification - straight pass-through.
3 · Conversion rate dips the week after a site release. Which team gets the first call?
Right lever, right owner: conversion belongs to product and UX, and a dip timed to a release points at the release. More traffic would paper over a leak instead of fixing it.