Beyond awareness: building full funnel growth frameworks that do not leak

A full funnel is a chain, and a chain loses something at every link

Most growth plans I am shown are plans for the top of the funnel. More reach, more clicks. Then a surprising amount of the money quietly falls out somewhere between the first visit and the second order.

A full funnel framework is simply the habit of looking at the whole journey at once. I use five stages:

1. Awareness: the right people learn that you exist.
2. Consideration: some of them look closely at what you sell.
3. Conversion: some of those buy for the first time.
4. Retention: some of the buyers come back and buy again.
5. Advocacy: some of the regulars bring other people with them.

The share of people who do not make it from one stage to the next is a leak. Every funnel leaks. The question is which leak costs the most, and which is the least expensive to fix.

Each box is a group of people, and each arrow is the share of one group that makes it into the next. A funnel that ends in advocacy refills itself.

If the idea of a funnel as counts and ratios is new to you, the article on the AIDA model rewired builds it from scratch, and the one on pirate metrics shows a close cousin of these five stages built for software products.

Worked example: a candle shop in Cork

Throughout this article I use an invented shop, Shandon Candle Co., an online candle business in Cork run by Aoife at shandoncandles.example. Nobody real, every number made up, so that every step can be shown.

In a typical month Aoife spends €3,000 on advertising. Here is her month, stage by stage:

• Awareness: her adverts and posts reach 200,000 people.
• Consideration: 3,000 of them land on a product page.
• Conversion: 240 put a candle in the basket, and 60 complete an order.
• Retention: 30% of new customers order again the following year.
• Advocacy: for every 100 new customers, five bring a friend who buys.

Her average order is €40, and half of that is margin after the cost of the candle, the box and the payment fee. So each order earns her €20. Her customers order about one and a half times a year while they stay, which is €30 of margin per customer per year.

On paper the month looks grand. In fact she is losing money on every new customer, and the rest of this article finds out where.

Give every stage one count you can collect honestly

A stage you cannot count is a stage you cannot fix. So the first job is dull and essential: one number per stage, defined the same way every month, collected with consent.

The definitions I use for a shop:

• Awareness: unique people reached, from the advertising platforms and your social accounts.
• Consideration: people who viewed at least one product page.
• Conversion: first orders, from the shop's own database, not from an analytics tool.
• Retention: the share of last year's new customers who ordered again this year.
• Advocacy: new customers who arrived through a referral code or named a customer who sent them.

Two traps are worth knowing about. The first is bounce rate. In Google Analytics 4 a session counts as engaged if it lasts longer than 10 seconds, includes a key event, or has two or more page views, and bounce rate is simply the share of sessions that were not engaged.¹ That is not the bounce rate of the older tool, so old rules of thumb like "above 70% is bad" do not carry over. I compare a page against itself over time.

The second trap is the urge to watch everything. You do not need screen recordings, heatmaps or scroll tracking. Five consented counts answer every question in this article, and five numbers get read, while a dashboard of 50 does not.

Worked example: where Aoife's five numbers come from

Reach comes from her advertising accounts: 200,000. Product page viewers come from her analytics, consented visitors scaled up by her known consent rate: 3,000. First orders come straight from her shop's order table, where a first order is any order from an email address never seen before: 60. Retention comes from the same table, a year later: of last year's 720 new customers, 216 ordered again, which is 30%. Referrals come from the referral code field at checkout.

Each definition is written down, so next month means the same as this month.

Customers are a product of rates, so every leak costs the same

Once each stage has a count, the links between them are rates. A rate is one count divided by the one before it. And the number of new customers is the top count multiplied by every rate on the way down.

$$C = R \times a \times b \times c$$

Here $C$ is the number of new customers in a month, $R$ is the number of people reached, $a$ is the share of those people who view a product page, $b$ is the share of product viewers who add something to the basket, and $c$ is the share of baskets that become an order. Each of $a$, $b$ and $c$ is a simple ratio of two counts. Foundations: ratios and rates

Because it is a multiplication, a 20% improvement anywhere in the chain gives 20% more customers. A better checkout is worth as many customers as 20% more reach. The difference is the price: more reach costs more every month, a better checkout is paid for once.

Better still, improvements multiply. Three improvements of 20% each give $1.2 \times 1.2 \times 1.2 = 1.728$, which is 72.8% more customers, not 60%. Foundations: percentages

Worked example: Aoife's chain in numbers

Aoife's rates, from her counts:

$$a = \frac{3{,}000}{200{,}000} = 1.5\%, \qquad b = \frac{240}{3{,}000} = 8\%, \qquad c = \frac{60}{240} = 25\%$$

Here $a$ is her product view rate, $b$ her basket rate and $c$ her checkout rate, each one count divided by the count before it, using the figures from her month. Foundations: ratios and rates

Putting them back together: $200{,}000 \times 0.015 \times 0.08 \times 0.25 = 60$ customers.

Her cost per new customer is $€3{,}000 / 60 = €50$. Keep that €50 in mind.

A customer is worth more than the first order, and that changes everything

The top three stages produce customers. The bottom two decide what a customer is worth.

The value of a new customer is the margin they bring this year plus the margin they bring in later years, with later money counted for a little less, because a euro next year is worth less than a euro today.

$$V = M + M \times \frac{r}{1 + d - r}$$

Here $V$ is the value of a new customer in today's money, $M$ is the margin an active customer brings in a year, $r$ is the retention rate (the share of customers who stay from one year to the next), and $d$ is the discount rate, the return you could get on the money elsewhere. The second term is the standard lifetime value formula for the years after the first. Foundations: lifetime value and Foundations: discounting

The old shortcut, average revenue times average lifespan times margin, is the same idea without the discount. It works when retention is low and flatters badly when retention is high, because it counts a euro in year five as a euro today. Always feed it margin, never revenue.

Worked example: what Aoife's customer is worth

Aoife's numbers: $M = €30$, $r = 0.30$, $d = 0.10$.

$$V = 30 + 30 \times \frac{0.30}{1 + 0.10 - 0.30} = 30 + 30 \times \frac{0.30}{0.80} = 30 + 11.25 = €41.25$$

Here $V$ is the value of one of Aoife's new customers, $M$ is the €30 a customer brings in a year, $r$ is her 30% retention rate and $d$ is the 10% she uses as her discount rate. Foundations: lifetime value

So each new customer is worth €41.25 and costs €50 to win. She loses €8.75 on every one. Across 60 customers a month that is $60 \times 8.75 = €525$ lost every month, by a shop that looks busy. More advertising would make it worse.

Find the leak by working out what each fix is worth

Staring at a funnel does not tell you where the leak is, because every rate sounds low. What tells you is one question per link: if I improved this to a rate I believe is achievable, how many euro is that worth?

"Achievable" comes from your own history (the best month you had), from a test you ran, or from a careful look at what is plainly broken. It never comes from an industry average quoted without a source.

Worked example: four fixes for Aoife

Aoife writes down one realistic improvement per link, and works out the monthly value of the customers each one produces. The baseline is 60 customers at €41.25, which is €2,475 of customer value a month for €3,000 of advertising.

1. Better targeting lifts the product view rate from 1.5% to 1.8%. Product viewers go from 3,000 to 3,600, customers from 60 to 72. Value: $72 \times 41.25 = €2{,}970$, a gain of €495.
2. Better product pages lift the basket rate from 8% to 10%. Baskets go from 240 to 300, customers from 60 to 75. Value: $75 \times 41.25 = €3{,}093.75$, a gain of €618.75.
3. A simpler checkout lifts the checkout rate from 25% to 30%. Customers go from 60 to 72. Value: €2,970, a gain of €495.
4. A post purchase programme lifts retention from 30% to 40%. Customers stay at 60, but each is now worth $30 + 30 \times 0.40 / 0.70 = €47.14$ (more exactly €47.142857). Value: $60 \times 47.142857 = €2{,}828.57$, a gain of €353.57.

All four together: $3{,}600 \times 0.10 \times 0.30 = 108$ customers, each worth €47.14, which is $108 \times 47.142857 = €5{,}091.43$, a gain of €2,616.43. That is far more than the four gains added up (€1,962.32), because the fixes multiply.

The horizontal axis lists the fixes, one bar each, and the last bar is all four at once. The vertical axis is the gain in euro of customer value per month, on top of the €2,475 Aoife has today. Two things stand out. The product pages are the biggest single leak. And the last bar is much taller than the four before it added together, which is the whole argument for a full funnel: fixes at different stages multiply each other.

Same €3,000 budget. Before, €2,475 of customer value and a loss of €525 a month. After, €5,091.43 and a surplus of €2,091.43. Her cost per new customer falls from €50 to $3{,}000 / 108 = €27.78$. Nothing at the top changed except the targeting.

The retention bar looks the smallest. Do not be fooled. It is the only fix that changes whether her advertising pays at all, as the section on retention shows.

Awareness: reach the right people, and do not starve the stage either

The old version of this article said businesses spend too much on awareness. That is half true. Reaching the wrong people is waste. Reaching the right people before they are ready to buy is not.

The best known evidence here comes from Les Binet and Peter Field, whose 2013 report for the Institute of Practitioners in Advertising found that the best balance across the campaigns they studied was roughly 60% of budget on long term brand building and 40% on short term activation.² That ratio comes from large advertisers, and I would not copy it to a candle shop in Cork. The lesson travels, though: people who remember your name next month are not a leak just because they did not buy this week.

So the question is not how much reach, but reach among whom, and at what cost per person who goes on to look closely.

Worked example: two audiences, one budget

Aoife splits her €3,000 evenly across two audiences. Audience one is a broad home and interiors audience: €1,500 reaches 120,000 people, and 1,200 of them view a product, a rate of 1%. Audience two is people interested in gifts and candles in Cork, Kerry and Limerick: €1,500 reaches 80,000 people, and 1,800 view a product, a rate of 2.25%.

Together: 200,000 reached, 3,000 viewers, 1.5%.

The cost per product viewer is $1{,}500 / 1{,}200 = €1.25$ for the broad audience and $1{,}500 / 1{,}800 = €0.83$ for the focused one. She does not switch the broad audience off, because some of it is the brand building that pays later. She tightens it: new creative that shows the candles lit in a real room, and the placements that sent no viewers at all excluded. The broad audience's view rate rises from 1% to 1.5%, so it now sends 1,800 viewers. Total viewers: 3,600, which is the 1.8% from fix one.

Consideration: answer the questions people actually have

A person on a product page is interested. Whether they put something in the basket depends on whether the page answers the questions in their head before they have to ask. For a candle, those questions are boringly practical. How long does it burn? How strong is the scent in a normal room? What does delivery cost, and when will it come? Can I send it straight to my sister in Galway as a gift?

Most product pages I review answer the questions the business finds interesting (the wax, the story) and skip the ones the buyer has. The fix is rarely clever: listen, then write the answer on the page, near the button.

How do you find the questions without watching people? You ask them. A one line question in the order confirmation email ("what nearly stopped you buying today?") and a read through the customer service inbox will give you the top five in a week, and a good few of them will surprise you. The article on the clarity framework goes further into removing confusion from a page.

Worked example: Aoife's product page

Aoife reads three months of customer emails and her order confirmation replies. The same four questions come up again and again: burn time, scent strength, delivery cost and gift wrapping. Her product pages answer none of them above the fold. They open with a paragraph about soy wax.

She adds four short lines next to the basket button: "Burns for about 45 hours", "Fills a living room, not a hall", "Free delivery over €35, otherwise €3.95, two working days with An Post", "Gift wrapped and sent with your note, free". Nothing else changes.

Over the next eight weeks the basket rate on those pages rises from 8% to 10%, which is fix two. Before trusting it, she checks she has enough baskets to tell a real change from noise, the subject of the article on statistical testing.

Conversion: at the last step, friction beats motivation

Someone with a full basket wants the thing. If they leave, the reason is usually friction. The Baymard Institute, who have collected cart abandonment studies for years, put the average documented abandonment rate at 70.22% across 50 studies.³ So Aoife's 75% is ordinary, which is not the same as fine.

The same research lists why shoppers abandon. The top reasons are extra costs that were too high (shipping, tax, fees) at 40%, delivery that was too slow at 20%, not trusting the site with card details at 19%, being asked to create an account at 18%, and a checkout that was too long or complicated at 17%.³ Nearly all of them are in the shop's control.

A useful way to think about it comes from BJ Fogg at Stanford, whose behaviour model says a behaviour happens when motivation, ability and a prompt come together at the same moment, and when it does not happen, at least one of the three is missing.⁴ At the checkout, motivation is already there. What is missing is ability: the surprise delivery charge, the password, the form that will not accept an Eircode with a space in it.

The old version of this article offered a formula, motivation times clarity divided by friction. I have taken it out: it was my shorthand, not a measured relationship. Fogg's model makes the same point without pretending to be arithmetic.

Speed belongs here too. The old article said a page should be interactive within three seconds. That metric, Time to Interactive, was removed from Google's Lighthouse tool in version 10.⁵ Google's current Core Web Vitals ask for the main content to load within 2.5 seconds, a response to a tap or click within 200 milliseconds, and a layout shift score of 0.1 or less, each measured at the 75th percentile of real visits on mobile and desktop.⁶ The article on Core Web Vitals explains each one.

Worked example: Aoife's checkout

Aoife walks through her own checkout on her phone and notes every moment she had to stop and think. Three stand out. Delivery cost first appears on the last page. The checkout insists on creating an account. The phone number field is mandatory and rejects numbers with spaces.

She fixes all three. Delivery cost appears in the basket as well as on the product page. Guest checkout becomes the default, with an optional password afterwards. The phone number becomes optional and accepts any format.

Her baskets stay at 240 a month. Completed orders rise from 60 to 72, so the checkout rate goes from 25% to 30%. That is fix three, worth €495 a month, paid for once.

Retention: the stage that decides whether growth compounds

Here is the uncomfortable fact from Aoife's numbers. At 30% retention, a customer is worth €41.25 and costs €50. No amount of awareness spending fixes that. It just loses money faster.

Retention is the multiplier on everything above it. And retention work is usually inexpensive compared with advertising, because these people already know you and gave you permission to write to them.

The horizontal axis is the share of customers who come back the following year. The vertical axis is euro. The rising line is what a new customer is worth at each retention rate, using the formula from earlier. The flat line is the €50 Aoife pays to win a customer today. Below about 44% retention her advertising loses money on every customer, above it every customer makes money, and the curve gets steeper the further right you go. Retention gains are worth more the better your retention already is.

Where does 44% come from? Setting the value equal to the cost and solving for $r$:

$$\frac{M(1 + d)}{1 + d - r} = 50 \quad\Rightarrow\quad 1 + d - r = \frac{30 \times 1.10}{50} = 0.66 \quad\Rightarrow\quad r = 1.10 - 0.66 = 0.44$$

Here $M$ is the €30 yearly margin, $d$ the 10% discount rate and $r$ the retention rate at which a customer is worth exactly the €50 it costs to win one. The first expression is the value formula from earlier with its two terms combined. Foundations: lifetime value

The articles on retention and referral and habit loops cover building the programme, and survival analysis covers measuring it.

Worked example: Aoife's first year programme

Aoife notices that most of her repeat customers buy their second candle within ten weeks of the first, when the first one runs out. So she builds three emails for new customers who agreed to hear from her: a care guide on day three (trim the wick, the first burn should reach the edge), a short note at week six with the matching refill, and a reminder at week nine with a small saving on a second candle.

A year on, 40% of that year's new customers have ordered again, up from 30%. That is fix four. Each new customer is now worth €47.14. With fixes one to three in place too, she wins them at €27.78 each, and every customer makes her money.

Advocacy: referrals are designed, and reviews have rules

Advocacy is where a funnel turns into a loop: a customer who brings a friend produces a new customer without an advertising bill. It rarely happens on its own at any useful scale. It happens when you ask at the right moment and make it easy.

Referred customers are worth having. A study of about 10,000 customers of a German bank, tracked for almost three years, found that referred customers had a higher margin, a higher retention rate, and an average value at least 16% above comparable customers who were not referred.⁷ That is a bank, not a candle shop, so I would not copy the 16% into a forecast. The direction is what matters.

Reviews are the other half of advocacy, and in Ireland they come with hard rules, set at EU level. The EU's Omnibus Directive added two practices to the list that is banned in all circumstances: claiming that reviews come from people who actually bought or used the product without taking reasonable and proportionate steps to check, and submitting or commissioning false reviews, or misrepresenting reviews, to promote a product. It also made it material information whether and how a business checks that its reviews come from real buyers.⁸ Ireland wrote both into the Consumer Protection Act 2007 through the Consumer Rights Act 2022, in force since the end of November 2022,⁹ and the Competition and Consumer Protection Commission (CCPC) enforces them.¹⁰ So you may ask every customer for a review. You may not write, buy or commission fake ones, or present a review as something it is not, and if you tell readers your reviews come from real customers, you need to have checked and to say how.

The new customers from advocacy add to the customers from advertising:

$$C_{\text{total}} = C \times (1 + k)$$

Here $C_{\text{total}}$ is all new customers in a month, $C$ is the new customers from advertising, and $k$ is the number of referred customers each new customer brings, on average, in their first year. Foundations: ratios and rates The article on referral and viral mechanics covers what happens when $k$ gets large.

Worked example: Aoife's referral card

Today $k = 0.05$: five referred customers for every 100 new ones, so $60 \times 0.05 = 3$ a month. Aoife adds a printed card to every parcel and a line in the week six email: give a friend €5 off their first candle and get €5 off your next one. The ask arrives when the customer has just lit the candle and likes it, which is the moment of delight the old version of this article rightly pointed at.

Referrals double to $k = 0.10$. On her original 60 customers that is $60 \times 0.10 = 6$ referred customers a month. Each costs her €10 in vouchers, so $6 \times 10 = €60$ a month, against $6 \times 50 = €300$ to win the same six through advertising. Each is worth at least the €41.25 of an ordinary customer, so the six bring $6 \times 41.25 = €247.50$ of value for €60.

For reviews, she sends one email after the second week asking for an honest review, with no reward attached. It is simpler to stay well clear of the line.

Fix one leak at a time, in order of value

A full funnel framework is not five projects at once. It is a short loop, run again and again, one leak at a time.

The steps run top to bottom. The decision is the honest part: when every link is in good shape, and only then, more awareness is the right next euro. One change at a time also tells you which change did the work.

Worked example: Aoife's quarter

Aoife orders the work by value per euro of effort. The product page lines come first: an afternoon of writing and the largest single gain. The targeting change is a week of her own time, so it comes second. The checkout fixes need her developer, so they come third, paid for once. The retention emails go live in the same quarter, but she can only judge them a year later.

Twelve months on, the same €3,000 a month of advertising wins 108 customers instead of 60, each worth €47.14 instead of €41.25, at €27.78 each instead of €50. A shop that lost €525 a month on its marketing now gains €2,091.43. Nothing about the top of the funnel got bigger. The funnel just stopped leaking.

That is the whole framework: count each stage, value each leak, fix the most valuable one, and remember that the bottom of the funnel decides whether the top is worth paying for. The article on the growth flywheel picks up where this one ends.

Footnotes

1. Google, [GA4] Engagement rate and bounce rate, Analytics Help, 2025, https://support.google.com/analytics/answer/12195621

2. Alison Hoad, The next chapter for The Long and The Short of It, Institute of Practitioners in Advertising (IPA), 13 September 2023, https://ipa.co.uk/knowledge/ipa-blog/the-next-chapter-for-the-long-and-the-short-of-it/

3. Richard Lam, Cart abandonment rate: 50 statistics and why people abandon, Baymard Institute, updated October 2026, https://baymard.com/blog/cart-abandonment-rate

4. BJ Fogg, Fogg Behavior Model, Behavior Design Lab, Stanford University, https://behaviormodel.org/

5. Google, Time to Interactive, Chrome for Developers, Lighthouse documentation, https://developer.chrome.com/docs/lighthouse/performance/interactive

6. Google, Web Vitals, web.dev, https://web.dev/articles/vitals

7. Philipp Schmitt, Bernd Skiera and Christophe Van den Bulte, Referral programs and customer value, Journal of Marketing, volume 75, issue 1, pages 46 to 59, American Marketing Association, 2011, https://faculty.wharton.upenn.edu/wp-content/uploads/2012/04/Schmitt-Skiera-vandenBulte-2011-Referral-Programs-Customer-Value.pdf

8. European Parliament and Council, Directive (EU) 2019/2161 (the Omnibus Directive), Article 3, inserting Article 7(6) and Annex I points 23b and 23c into Directive 2005/29/EC, Official Journal of the European Union, 2019, https://eur-lex.europa.eu/eli/dir/2019/2161/oj/eng

9. Law Reform Commission, Consumer Protection Act 2007, revised, sections 46(3B) and 55(1)(za) and (zb), as inserted by the Consumer Rights Act 2022, Revised Acts, https://revisedacts.lawreform.ie/eli/2007/act/19/revised/en/html

10. Competition and Consumer Protection Commission, CCPC welcomes new rights for consumers, CCPC, 28 November 2022, https://www.ccpc.ie/news-and-media/news/article/2022/11/28/ccpc-welcomes-new-rights-for-consumers