Tabellab

Landing Page Scorecard: Bounce Rate, Conversion Rate, Revenue per Session and the Engaged Conversion Rate

Lesson 1 of 20Lab

All lessons
  1. Landing Page Scorecard: Bounce Rate, Conversion Rate, Revenue per Session and the Engaged Conversion RateLabFree
  2. Message Match: Scoring Ad-to-Page Consistency and Comparing Conversion Rates by Landing Page TypeLab Locked
  3. Above-the-Fold Audit: a First-Screen Checklist with SUM, AND, IFS, Pass Shares and a Formatting RuleLab Locked
  4. Form Length vs Conversion: Start Rate, Completion Rate, Absolute References and the Cost of One More Field with SLOPELab Locked
  5. Page Speed Impact: Bounce and Conversion Rate by Load-Time Band and the Orders Missed Against the Fastest BandLab Locked
  6. Valuing a Speed Fix: Seconds to Save with MAX, a Relative Conversion Gain, Extra Revenue per Page and a Ranking with RANK.EQLab Locked
  7. Scroll Depth: Reach per Scroll Mark, Step Losses and the Share of Visitors Who See Each Page SectionLab Locked
  8. Bounce vs Conversion by Traffic Source: SUMIFS Roll-Ups, Reading the Two Rates Together and a Mismatch FlagLab Locked
  9. Device Segments: a Page-by-Device Table with Two-Criteria SUMIFS and Mixed References, the Mobile-to-Desktop Ratio and the Gap in OrdersLab Locked
  10. Form Length and Lead Quality: Cost per Lead, Qualified Rate, Cost per Qualified Lead and the Best Variant with MIN and XLOOKUPLab Locked
  11. Variant Comparison: Control and Variant Rates, Relative Uplift vs Percentage Points and a First Reading with MIN and IFSLab Locked
  12. Is the Difference Real? Pooled Rate, Standard Error, z-Score and p-Value for Two Conversion RatesLab Locked
  13. Planning a Test: Sample Size by Rule of Thumb, Days Needed with ROUNDUP, Full Weeks with CEILING and an End DateLab Locked
  14. Test Results by Traffic Source: Segment Rates, a Split Check and Why an Overall Uplift Can MisleadLab Locked
  15. The Page Funnel: Step Rates, an Index Against a Benchmark with a Row-Locked Reference and the Weakest Step by NameLab Locked
  16. Speed Over Time: Correlation, Slope and a Conversion Forecast at a Target Load Time with FORECAST.LINEAR and a Scatter ChartLab Locked
  17. What-If Grid for a Redesign: Extra Contribution by Traffic and Conversion Rate with Mixed References, Payback and a Heat MapLab Locked
  18. Prioritising the Backlog: Expected Extra Orders, Confidence, Effort, a Ranking and the Top Three with LARGELab Locked
  19. Quality Checks Before Reporting: Duplicates with COUNTIF, Impossible Values with OR, Untagged Visits and a Clean Conversion RateLab Locked
  20. Final Project: Landing Page Dashboard with SUMIFS Roll-Ups, Bounce and Conversion Rate, Revenue per Session, a Status and a Combo ChartLab Locked

You are not signed in — the lab works, but your progress is not saved. Sign up

Lesson 1 of 20 is free. Unlock all 20 hands-on labs, the final exam and your certificate.

Buy course — $9.90
Loading the spreadsheet…

Why every landing page needs a scorecard

A landing page is the first page a visitor sees after clicking an ad, a link in an email, a pin or a search result. It has one job: to turn that click into the next step, usually an order or an enquiry. Most businesses run many landing pages at once, and they almost never perform equally. Some quietly earn most of the revenue. Others receive a large share of the paid traffic and return very little. The trouble is that you cannot see the difference in the raw numbers.

An analytics export typically lists sessions, bounces, orders and revenue per page. Sorted by orders, the busiest pages always come out on top, simply because they are busy. Landing page optimisation starts when you stop asking how many and start asking what share: of the people who arrived, how many left at once, how many bought, and how much did each visit earn? A scorecard answers those questions with four simple rates and puts every page on the same footing, whether it had one thousand sessions or fifteen thousand.

This free lesson builds such a scorecard in a spreadsheet. The example is Pebblecroft Home, a fictional online shop for home-office furniture and lighting, with ten landing pages that received paid and email traffic in March 2026. The company, the pages and all numbers are invented for learning, and any benchmark mentioned is illustrative. You work with exactly the same data in the lab above, where each formula is checked as you type it.

The raw material: sessions, bounces, orders and revenue

Four columns are enough to start. It is worth being precise about what each one counts, because the definitions differ slightly between analytics tools and the rates inherit those definitions.

  • Sessions are visits. One person who arrives on Monday and again on Thursday counts as two sessions. For landing pages you count the sessions that started on the page, not all views of it.
  • Bounces are sessions that ended without a second page view or any recorded interaction. Some tools define the opposite, an "engaged session", as a visit that lasted longer than a few seconds, viewed a second page or triggered a conversion. A bounce is then a session that was not engaged.
  • Orders are completed purchases in sessions that started on the page. On a lead-generation page the same column would hold form submissions; the arithmetic is identical.
  • Revenue is the value of those orders. Use the same basis for every page, for example revenue after discounts and before delivery charges.
Check how your own tool defines a bounce before you compare your numbers with anyone else's. A site that counts a ten-second visit as engaged will report a much lower bounce rate than one that requires a second page view, with identical visitor behaviour.

Four rates that make pages comparable

Each of the four rates divides one column by another. They answer four different questions, and a page can look good on one and poor on the next.

The bounce rate is bounces ÷ sessions. It measures the first few seconds: did the visitor find what the ad or link promised, quickly enough to stay? A high bounce rate points to the top of the page (headline, image, loading time) or to the traffic itself, when the wrong people are being sent.

The conversion rate is orders ÷ sessions. It is the headline number of landing page optimisation, because it measures the whole job of the page from arrival to order. Its weakness is that it treats every order as equal. A page that sells desk lamps for €75 and a page that sells office fit-outs for €1,300 cannot be compared on conversion rate alone.

Revenue per session is revenue ÷ sessions. It equals conversion rate × average order value, so it corrects exactly that weakness: it tells you what one visit to the page is worth. That is also the number to hold against the cost of a click. If a visit earns €1.40 in revenue and the click costs €0.90, there is little left once the goods are paid for.

The engaged conversion rate is orders ÷ (sessions − bounces). It removes the visitors who left at once and asks how well the page persuades those who stayed. Together with the bounce rate it splits the conversion rate into two stages: getting people to stay, and getting those who stay to buy. The two stages have different causes and different fixes, which is why it pays to look at them separately.

From session to order: all ten landing pages, March 202651,565sessions27,692bounces23,873engaged sessions1,125ordersbounce rate 53.7 %46.3 % stayengaged conversionrate 4.71 %conversion rate = orders ÷ sessions = 2.18 %
Figure 1: The same 1,125 orders give two different rates depending on the divisor: 2.18 % of all sessions, 4.71 % of the sessions that did not bounce.

Figure 1 shows the two stages for all ten pages together. Of 51,565 sessions, 27,692 bounced. The remaining 23,873 engaged sessions produced 1,125 orders. Divide the orders by all sessions and you get the conversion rate of 2.18 %; divide them by the engaged sessions only and you get 4.71 %. Neither number is more correct than the other. They have different divisors and answer different questions.

A fully worked example: Pebblecroft Home in March 2026

Here is the raw data of the ten landing pages, exactly as it appears in columns A to E of the lab sheet (rows 4 to 13, with the totals in row 14).

Landing pageSessionsBouncesOrdersRevenue €
/desks/standing-desk-sale8,4203,705244€87,840
/chairs/ergonomic-chairs6,1502,952160€44,800
/lighting/desk-lamps4,3802,54079€5,925
/bundles/home-office-starter3,2601,174137€61,650
/storage/shelving2,9401,82338€4,940
/desks/compact-desks5,0752,335132€36,960
/accessories/monitor-arms3,6101,87883€7,055
/chairs/kids-study-chairs1,8901,15321€2,730
/sale/spring-clearance14,6009,636190€19,950
/b2b/team-workspaces1,24049641€53,300
All pages51,56527,6921,125€325,150

A first look at the raw numbers suggests an obvious ranking. /sale/spring-clearance has by far the most sessions (14,600) and the second-highest number of orders (190), so it looks like one of the best pages. /b2b/team-workspaces has the fewest sessions and only 41 orders, so it looks like the weakest. Both impressions are wrong, as the rates will show.

Step by step: the formulas

  1. Bounce rate (F4): type =C4/B4, bounces divided by sessions. For the first page that is 3,705 ÷ 8,420 = 44.0 %.
  2. Conversion rate (G4): =D4/B4, orders divided by sessions: 244 ÷ 8,420 = 2.90 %.
  3. Revenue per session (H4): =E4/B4, revenue divided by sessions: €87,840 ÷ 8,420 = €10.43.
  4. Engaged conversion rate (I4): =D4/(B4-C4), orders divided by the sessions that did not bounce: 244 ÷ (8,420 − 3,705) = 244 ÷ 4,715 = 5.17 %. The brackets matter. Without them the spreadsheet divides first and subtracts afterwards.
  5. Select F4:I4 and copy the formulas down to row 13. The references are relative, so row 5 automatically uses C5, B5 and so on.
  6. Overall conversion rate (G14): =SUM(D4:D13)/SUM(B4:B13). SUM adds all the numbers in a range; the formula divides total orders by total sessions.
Format rate cells as percentages with one or two decimals. The cell still holds the decimal number (0.029); the format only changes how it is shown (2.90 %). Multiplying the formula by 100 as well is a classic source of results that are off by a factor of one hundred.

Copied down, the four formulas produce this scorecard:

Landing pageBounce rateConversion rateRevenue per sessionEngaged conversion rate
/desks/standing-desk-sale44.0 %2.90 %€10.435.17 %
/chairs/ergonomic-chairs48.0 %2.60 %€7.285.00 %
/lighting/desk-lamps58.0 %1.80 %€1.354.29 %
/bundles/home-office-starter36.0 %4.20 %€18.916.57 %
/storage/shelving62.0 %1.29 %€1.683.40 %
/desks/compact-desks46.0 %2.60 %€7.284.82 %
/accessories/monitor-arms52.0 %2.30 %€1.954.79 %
/chairs/kids-study-chairs61.0 %1.11 %€1.442.85 %
/sale/spring-clearance66.0 %1.30 %€1.373.83 %
/b2b/team-workspaces40.0 %3.31 %€42.985.51 %
All pages53.7 %2.18 %€6.314.71 %

Conversion rate by landing pageConversion rateAll pages together: 2.18 %4.20 %/bundles/home-office-starter3.31 %/b2b/team-workspaces2.90 %/desks/standing-desk-sale2.60 %/chairs/ergonomic-chairs2.60 %/desks/compact-desks2.30 %/accessories/monitor-arms1.80 %/lighting/desk-lamps1.30 %/sale/spring-clearance1.29 %/storage/shelving1.11 %/chairs/kids-study-chairs
Figure 2: Conversion rate per landing page, sorted. Six pages lie above the overall rate of 2.18 % and four below; the largest page by far, /sale/spring-clearance, is among the weakest.

Reading the scorecard

Start with the page that looked best in the raw data. /sale/spring-clearance converts at 1.30 %, well below the overall rate of 2.18 %, and 66.0 % of its visitors bounce, the highest share of all ten pages. Its 190 orders are the product of sheer traffic. A visit to it earns €1.37. If that traffic is paid for by the click, this page is the first candidate for a closer look: either the page does not deliver what the sale ads promise, or the ads reach people who were never going to buy.

Now the page that looked weakest. /b2b/team-workspaces has only 1,240 sessions, but 3.31 % of them convert, and because a workspace order is large, each session earns €42.98. That is more than thirty times the value of a clearance visit. On conversion rate the page ranks second; on revenue per session it is first by a wide margin. Whether it deserves more traffic depends on how many more such buyers exist, but the scorecard has at least made the question visible.

/bundles/home-office-starter is the most consistent performer: the lowest bounce rate (36.0 %), the highest conversion rate (4.20 %) and the highest engaged conversion rate (6.57 %). A page like this is a useful internal reference. Before you look for outside advice, compare your weak pages with your own best one: what does its first screen show, how long is it, what does the button say?

The engaged conversion rate earns its place when two pages have almost the same conversion rate for different reasons. Compare /storage/shelving and /sale/spring-clearance. Both convert at about 1.3 % (1.29 % and 1.30 %). The shelving page loses fewer people at the door, with a bounce rate of 62.0 % against 66.0 %, but persuades fewer of those who stay: 3.40 % against 3.83 %. For the clearance page the first screen and the incoming traffic are the place to start. For the shelving page the offer further down deserves at least as much attention.

At the bottom of the table, /chairs/kids-study-chairs has the lowest conversion rate (1.11 %) and the lowest engaged conversion rate (2.85 %). With only 21 orders, however, its rates rest on very few events. One good or bad week could move them noticeably. Treat the rates of small pages as rough indications and collect more data before drawing firm conclusions; later lessons show how to judge whether a difference is larger than chance.

The total row: never average the rates

The overall conversion rate is total orders ÷ total sessions: 1,125 ÷ 51,565 = 2.18 %. A tempting shortcut is to average the ten page rates in column G with =AVERAGE(G4:G13). That returns 2.34 %, which is not the rate of the site.

The reason is weighting. An average of rates gives every page one vote. /sale/spring-clearance, with 14,600 sessions and a rate of 1.30 %, counts exactly as much as /b2b/team-workspaces with 1,240 sessions and 3.31 %. In reality the clearance page accounts for 28 % of all sessions and pulls the true rate down. Dividing the sums weights each page by its traffic automatically. The same rule applies to the bounce rate (27,692 ÷ 51,565 = 53.7 %), to revenue per session (€325,150 ÷ 51,565 = €6.31) and to every rate you will meet in this course: build totals from the sums, not from the rates.

Common mistakes

  • Ranking pages by orders or revenue. Volume mostly reflects how much traffic a page was given. Rank by rates, then look at volume to judge how much a fix would be worth.
  • Dividing by the wrong column. Orders ÷ bounces or revenue ÷ orders are valid numbers (the second is the average order value), but they are not the rates asked for. Say the formula aloud: "orders per session".
  • Forgetting the brackets in =D4/(B4-C4). =D4/B4-C4 calculates a tiny rate and then subtracts thousands of bounces.
  • Averaging rates for the total row instead of dividing the sums.
  • Treating a high bounce rate as bad in itself. A visitor who arrives from a brand search, checks the delivery time and leaves satisfied is a bounce. Always read the bounce rate next to the conversion rate.
  • Comparing pages with different goals on conversion rate alone. Use revenue per session, or value per session for lead pages, when order values differ widely.
  • Reading small pages too precisely. A rate built on twenty orders is a rough estimate.

What a scorecard cannot tell you

A scorecard shows where to look, not why a page underperforms. A weak page may have a poor first screen, a slow load, a long form, or simply worse traffic than its neighbours. The rates also mix every visitor together: phone and desktop, search and social, new and returning. Two pages with the same conversion rate can hide very different stories once those groups are separated.

It also describes one period. March may include a sale, a newsletter or a week with a tracking fault. Before acting on a single month, compare it with the months before, and check that the data is sound. None of this makes the scorecard less useful. It is the map that tells you which page to open first, and the remaining lessons supply the tools for finding the cause.

Using the scorecard on your own pages

  1. Export sessions, bounces (or engaged sessions), conversions and revenue by landing page for a full month.
  2. Keep the pages that received campaign traffic and have enough sessions to say something, for example at least a few hundred.
  3. Add the four rate columns and the total row exactly as above.
  4. Sort by revenue per session and mark the pages that combine high traffic with a low value per visit. They are where an improvement pays most.
  5. For each marked page, compare bounce rate and engaged conversion rate with your best page to see which stage is weaker.
  6. Write down one hypothesis per page before you change anything, so that you can test it instead of guessing.

What the rest of the course covers

The scorecard is lesson 1 of 20. Every later lab uses a new data set from Pebblecroft Home and adds one technique for finding out why a page performs as it does and what a change would be worth.

  • Message match and the first screen (lessons 2 and 3): you score how well each page repeats the promise of its ad, compare dedicated pages with category pages and the homepage, and audit the first screen of twelve pages with a checklist.
  • Forms (lessons 4 and 10): you measure what each extra form field costs in submits and then weigh that against lead quality with cost per qualified lead.
  • Page speed (lessons 5, 6 and 16): you compare conversion by load-time band, value a speed fix page by page and follow one page through a twelve-week speed project with a forecast.
  • Scroll depth and the page funnel (lessons 7 and 15): you find out how many visitors ever reach each section and which step from page view to order is weakest.
  • Segments (lessons 8, 9 and 14): you break bounce and conversion down by traffic source and device, and learn why an uneven split can make a test result misleading.
  • Variant comparison (lessons 11 to 13): you compare control and variant, check significance step by step and plan how long a test must run.
  • Decisions and reporting (lessons 17 to 20): you build a what-if grid for a redesign, prioritise a backlog of ideas, check an export for faulty rows and finish with a landing page dashboard.

Summary

Four divisions turn a raw export into a landing page scorecard: bounce rate =C4/B4, conversion rate =D4/B4, revenue per session =E4/B4 and engaged conversion rate =D4/(B4-C4). The total row divides the sums, =SUM(D4:D13)/SUM(B4:B13), and never averages the page rates. At Pebblecroft Home the scorecard overturns the picture given by the raw counts: the page with the most traffic converts at 1.30 % and earns €1.37 per visit, while the smallest page earns €42.98.

Now scroll up to the lab and build the sheet yourself. It takes about fifteen minutes, each formula is checked as you enter it, and the layout works unchanged for an export of your own landing pages.

Frequently asked questions

How do I calculate the conversion rate of a landing page in Excel?

Divide the orders (or leads) by the sessions that started on the page, for example =D4/B4, and format the cell as a percentage. Copy the formula down for the other pages. For the total, divide the sum of orders by the sum of sessions with =SUM(D4:D13)/SUM(B4:B13).

What is the difference between bounce rate and conversion rate?

The bounce rate is the share of sessions that left without a second page view or interaction; it describes the first seconds of a visit. The conversion rate is the share of sessions that ended with an order or lead; it describes the whole visit. A page needs both: a low bounce rate with a low conversion rate loses visitors later, at the offer or the checkout.

What is an engaged conversion rate?

It is orders divided by the sessions that did not bounce: =D4/(B4-C4). It shows how well a page persuades the visitors who stayed, separately from how many left immediately. Comparing it with the bounce rate tells you which of the two stages needs work.

Why should I not average the conversion rates of my pages?

An average of rates gives every page the same weight, however little traffic it has. The true overall rate is total orders divided by total sessions, which weights each page by its sessions. In the worked example the average of the page rates is 2.34 % while the real overall rate is 2.18 %.

Is a high bounce rate always a problem?

No. Visitors who find what they need on the first page and leave are counted as bounces, and discovery channels naturally bounce more than brand search. A high bounce rate is a concern when it comes with a low conversion rate, which suggests a mismatch between the ad and the page or poorly targeted traffic.

What is revenue per session and why use it?

Revenue per session is revenue divided by sessions, which equals conversion rate times average order value. It lets you compare pages that sell items at very different prices and can be set directly against the cost of a click to see whether the traffic pays for itself.

Keep going: Landing Page Optimisation: Message Match, Above-the-Fold Checks, Form Length, Page Speed, Scroll Depth & Variant Comparison

This free lab is lesson 1 of 20. Unlock the other 19 hands-on labs, the final exam and your certificate for a one-time $9.90.

  1. Message Match: Scoring Ad-to-Page Consistency and Comparing Conversion Rates by Landing Page Type
  2. Above-the-Fold Audit: a First-Screen Checklist with SUM, AND, IFS, Pass Shares and a Formatting Rule
  3. Form Length vs Conversion: Start Rate, Completion Rate, Absolute References and the Cost of One More Field with SLOPE
  4. Page Speed Impact: Bounce and Conversion Rate by Load-Time Band and the Orders Missed Against the Fastest Band
  5. Valuing a Speed Fix: Seconds to Save with MAX, a Relative Conversion Gain, Extra Revenue per Page and a Ranking with RANK.EQ
  6. Scroll Depth: Reach per Scroll Mark, Step Losses and the Share of Visitors Who See Each Page Section
  7. Bounce vs Conversion by Traffic Source: SUMIFS Roll-Ups, Reading the Two Rates Together and a Mismatch Flag
  8. Device Segments: a Page-by-Device Table with Two-Criteria SUMIFS and Mixed References, the Mobile-to-Desktop Ratio and the Gap in Orders
  9. Form Length and Lead Quality: Cost per Lead, Qualified Rate, Cost per Qualified Lead and the Best Variant with MIN and XLOOKUP
  10. Variant Comparison: Control and Variant Rates, Relative Uplift vs Percentage Points and a First Reading with MIN and IFS
  11. Is the Difference Real? Pooled Rate, Standard Error, z-Score and p-Value for Two Conversion Rates
  12. Planning a Test: Sample Size by Rule of Thumb, Days Needed with ROUNDUP, Full Weeks with CEILING and an End Date
  13. Test Results by Traffic Source: Segment Rates, a Split Check and Why an Overall Uplift Can Mislead
  14. The Page Funnel: Step Rates, an Index Against a Benchmark with a Row-Locked Reference and the Weakest Step by Name
  15. Speed Over Time: Correlation, Slope and a Conversion Forecast at a Target Load Time with FORECAST.LINEAR and a Scatter Chart
  16. What-If Grid for a Redesign: Extra Contribution by Traffic and Conversion Rate with Mixed References, Payback and a Heat Map
  17. Prioritising the Backlog: Expected Extra Orders, Confidence, Effort, a Ranking and the Top Three with LARGE
  18. Quality Checks Before Reporting: Duplicates with COUNTIF, Impossible Values with OR, Untagged Visits and a Clean Conversion Rate
  19. Final Project: Landing Page Dashboard with SUMIFS Roll-Ups, Bounce and Conversion Rate, Revenue per Session, a Status and a Combo Chart

Buy course — $9.90 See the full course