The four gates, measured precisely 16 min live
In a3 the leader track learned to read the funnel's shape. You compute it. Each gate is a ratio: sessions that passed ÷ sessions that reached the gate. Four gates, four divisions, and their product is the session CR to the decimal - that multiplication is the whole reason funnel metrics exist.
LiveGate 1 - PDP penetration, the underrated one5 min▶
Definition: sessions with at least one view_item ÷ all sessions. Mango Lane: 144,000 ÷ 360,000 = 40%. Six in ten sessions never see a single product page - they bounce off the homepage, wander category listings, or lose a fight with search. Nobody adds to cart from a listing thumbnail; the PDP is where conversion begins, so this gate caps everything downstream.
- Owners: navigation, search, merchandising, landing-page relevance. Not the checkout team.
- Why it is underrated: most dashboards ship cart and checkout rates by default; PDP penetration you usually have to build. Which means most teams cannot see the gate that filters out 60% of their traffic.
LiveGates 2-4 - cart, checkout start, completion7 min▶
- Gate 2 · add-to-cart rate: cart sessions ÷ PDP sessions = 36,000 ÷ 144,000 = 25%. Of sessions that saw a product, a quarter put something in the cart. Owners: PDP content, price, reviews, size guides, imagery.
- Gate 3 · checkout start: checkout sessions ÷ cart sessions = 16,200 ÷ 36,000 = 45%. More than half of carts never reach checkout - browsing carts, shipping-cost shock, "saving it for payday". Owners: cart UX, shipping transparency, cart reminders.
- Gate 4 · checkout completion: purchases ÷ checkout starts = 9,720 ÷ 16,200 = 60%. Of shoppers who started paying, six in ten finished. Owners: checkout form length, payment methods, errors, trust signals.
Self-studyWho counts sessions, who counts users3 min read▶
Today's gates are all session-based, matching Shopify's funnel report (sessions with cart additions → reached checkout → completed checkout). GA4's purchase journey report measures the same shape but counts users at each of its five steps (session_start → view_item → add_to_cart → begin_checkout → purchase), and it is a closed funnel by default - a shopper only counts at a step if they passed through the previous steps.
- Consequence: GA4's step-to-step percentages will not match your session-based gates even with perfect tracking. Different unit (users vs sessions), different funnel rules (closed vs open).
- The move: pick one framing for your own review pack (this course: sessions), and translate before comparing anything across tools. b8's spec sheet has a column for exactly this.
The denominator trap at every gate 12 min live
b1 warned you conversion rate has two denominators. Every gate has the same problem. Take add-to-cart: is Mango Lane's ATC rate 25% or 10%? Both - and if you quote one against a benchmark built on the other, you will diagnose a healthy store as sick, or a sick one as healthy.
| Framing | Formula | Mango Lane | When to use it |
|---|---|---|---|
| ATC of PDP sessions (gate framing) | cart sessions ÷ PDP sessions | 36,000 ÷ 144,000 = 25% | Diagnosing the PDP itself: content, price, reviews. Isolates gate 2 from gate 1. |
| ATC of ALL sessions (Littledata / Shopify framing) | cart sessions ÷ all sessions | 36,000 ÷ 360,000 = 10% | Comparing against session-based benchmarks (Littledata's Shopify average is 4.6%) and tracking overall shopping intent. |
LiveSame trap, checkout edition5 min▶
Checkout completion is 60% of checkout starts (9,720 ÷ 16,200). But 9,720 purchases ÷ 360,000 sessions is 2.7% of all sessions - the same purchases, a 22x difference in the headline number, purely from the denominator. Neither is wrong; they answer different questions ("how leaky is checkout?" vs "how much traffic becomes orders?").
- Mango Lane's session-based ATC of 10% is genuinely strong - more than double Littledata's 4.6% Shopify average. The 25% gate framing cannot be compared to that 4.6% at all.
- The habit: every funnel number you write down gets its denominator written next to it. Every benchmark you cite, find the source's denominator first. The homework makes you do exactly this for your own tools.
How this bites: a team reads "average add-to-cart rate is 4.6%" in a Shopify benchmark report, checks their own gate-framed ATC of 25%, and concludes they are 5x above average - champagne. The benchmark was session-based; their true session-based figure was 8%. Still good, but the celebration was built on a denominator mismatch, and next quarter someone reverses the error in the other direction and declares a crisis.
Benchmarks per gate, honestly cited 10 min live
Gate rates without yardsticks are just numbers. Here is each gate against the strongest public benchmark available - with the source and its denominator named, as promised.
LiveMango Lane vs the published numbers6 min▶
| Gate | Mango Lane | Benchmark | Verdict |
|---|---|---|---|
| Add-to-cart (session-based) | 10% | Dynamic Yield: ~1.7% (luxury) to ~10% (food/beauty) by vertical; Littledata Shopify benchmark, 2025: 4.6% avg | Strong - top of the published range |
| Cart → purchase | 9,720 ÷ 36,000 = 27% (73% abandonment) | Baymard Institute: 70.22% average cart abandonment (average of 50 studies, updated Sep 2025) | Right at benchmark - normal, not a crisis |
| Checkout completion | 60% | Littledata Shopify benchmark, 2025: avg 45%, top 20% 59%, top 10% 66% | Strong - around the top decile boundary |
| PDP penetration | 40% | No solid public benchmark exists - track your own trend instead | Watch the trend, not a yardstick |
Honesty note: Mango Lane is written as a deliberately healthy fashion store - its 2.7% session CR sits above IRP Commerce's fashion average of 1.70%, and its 60% completion beats Littledata's 45% average. Real stores you audit will usually sit below these numbers, and that is exactly what the gate audit is for.
Self-studySeeing WHERE checkout leaks - GA4's checkout journey4 min read▶
Gate 4 says checkout loses 40% of starters, but not where. GA4's checkout journey report zooms into the gate itself with four steps: begin_checkout → add_shipping_info → add_payment_info → purchase, showing an abandonment percentage at each step. A spike between shipping and payment points at shipping-cost shock; a spike after payment info points at payment errors or a scary final screen. It is the same funnel logic, one level down - gates inside a gate.
- Pair it with GA4's purchase journey report (the five-step user funnel from b1) for the full picture: journey report for the funnel, checkout report for the microscope.
- b5 uses exactly this drill-down move when you diagnose a real drop.
The gate audit drill ★ 7 min · pens out
A bad month lands on your desk: PDP penetration fell from 40% to 34%; all other gates held flat. Before anyone opens a dashboard, you can compute exactly what that cost and who owns the fix. This is the drill.
LiveCompute the damage, name the owner6 min▶
New PDP sessions: 360,000 x 0.34 = 122,400 (was 144,000 - the gate now filters out 21,600 more sessions).
Walk the flat gates: carts = 122,400 x 0.25 = 30,600; checkout starts = 30,600 x 0.45 = 13,770; orders = 13,770 x 0.60 = 8,262.
New GMV: 8,262 x $68 = $561,816. Versus baseline: 9,720 - 8,262 = 1,458 orders lost, $660,960 - $561,816 = $99,144 GMV lost - a 15% revenue hit from a six-point slip in one gate.
Name the owner: the leak is upstream of the PDP, so it belongs to navigation, search, and merchandising - homepage modules, search relevance, category pages, landing-page match. The checkout team is innocent this month; do not spend their sprint on it.
Self-studyRun the drill on the other gates4 min read▶
The same arithmetic works for any gate, and running it on all four builds your intuition for which slips hurt most. Try these on paper (baseline: 360,000 sessions, gates 0.40 / 0.25 / 0.45 / 0.60, AOV $68):
- Add-to-cart slips 25% → 21%: orders = 360,000 x 0.40 x 0.21 x 0.45 x 0.60 = 8,165. Lost: 1,555 orders, ~$105,740 GMV. Owner: PDP content, price, reviews.
- Completion slips 60% → 52%: orders = 16,200 x 0.52 = 8,424. Lost: 1,296 orders, ~$88,128 GMV. Owner: checkout and payments - and the GA4 checkout journey report tells you which step.
- The pattern: a proportional slip in ANY gate costs the same share of orders (the gates multiply, so a 10% relative drop anywhere is a 10% order drop). What differs is who owns the fix and how fast it can ship - which is why naming the gate matters more than sizing the loss.
Try it yourself - this week ◐ 25-35 min total
- Compute your store's four gates for last month: PDP penetration, add-to-cart, checkout start, checkout completion. Multiply them - do they reproduce your session CR?
- For each tool you use (GA4, Shopify, anything else), mark which denominator its version of each gate uses - of-all-sessions or of-the-previous-gate. Write it next to the number.
- Circle the gate furthest below its benchmark (use Part 3's table; for PDP penetration compare against your own previous months). That circle is your store's cheapest improvement.
- If you have GA4, open the checkout journey report and note which of the three checkout steps leaks most. You will want that answer in b5.
Sources this session draws on
Gate definitions come from the platforms; yardsticks come from published benchmark panels of very different rigor - each is labeled with what it actually measures. This page draws on:
Three questions before you go 🎯 ◐ 90 seconds
1 · Which of these is Mango Lane's PDP penetration?
PDP penetration counts SESSIONS that reached at least one product page, over all sessions. Counting raw view_item events would let one shopper browsing 30 dresses inflate the number; counting sessions asks the real question - did this visit ever reach a product at all?
2 · One report says Mango Lane's add-to-cart rate is 25%, another says 10%. What is going on?
Same numerator, two honest denominators. The gate framing (25%) isolates PDP performance; the session framing (10%) is what benchmarks like Littledata's 4.6% use. Quote a gate-framed number against a session-based benchmark and you will be wrong by 2.5x while every input is correct.
3 · A payment provider has an outage for a week. Which gate takes the hit, and what stays flat?
Shoppers still reach products, add to cart, and start checkout - the failure happens when they try to PAY, which is gate 4 (purchases ÷ checkout starts). One gate moving while the others hold is the classic signature that tells you where to look - GA4's checkout journey report would show the leak at the add_payment_info → purchase step.