# Marketing Funnel The marketing funnel is the drop from awareness toward advocacy, used to see where prospective customers leave. **The question.** Of the people who could buy, how many reach each stage, and where do they leave? Also known as: Also called the purchase funnel, sales funnel, or conversion funnel. ## When to use it - Growth has stalled and nobody can say whether the problem is reach, consideration, conversion, or retention. - Marketing reports spend and leads, sales reports closed deals, and nobody owns the middle. - You are comparing the same measures across channels, segments, or periods to find where cost per customer is highest. - A business-to-business pipeline needs stage definitions that marketing and sales both accept. ## When to avoid it - You want to explain how people decide. Buyers loop back, skip stages, and add brands while they shop. Use research on real buying behavior, and use the funnel only to measure. - Volume is too small for stage rates to mean anything, such as a dozen enterprise deals a year. A deal-by-deal review is more useful than percentages. - The business is mostly repeat purchase by existing customers. The upper stages matter little, and a retention view fits better. - The question is why people would choose the brand over others. That is brand equity, and the funnel only counts the outcome. ## Overview The marketing funnel counts how many prospective customers reach each stage on the way to a purchase, commonly awareness, consideration, intent, purchase, loyalty, and advocacy. It supports a diagnostic choice: which stage to work on, and which team owns it, based on where the rate between stages is lowest against an earlier period, another segment, or another channel. Use it as a measurement frame, not as a theory of how people decide. Buyers loop back, skip stages, and add brands as they shop, which is why McKinsey proposed a circular model in 2009. Stage definitions have to be countable, and sales and marketing have to agree on them, or the rates will measure the disagreement. ## The stages ### Awareness The prospect knows the brand or offer exists. Measured by survey (unaided and aided recall), reach, or branded search. ### Consideration The prospect includes the brand when comparing options. Also called interest. Measured by shortlist inclusion in surveys, content engagement, or requests for information. A brand outside the initial set can still enter later, because shoppers add options as they evaluate. ### Intent The prospect signals a decision is coming, for example a quote request, a trial, a demo, or a basket with items in it. The data is behavioral and is usually the easiest stage to count in digital channels. ### Purchase The prospect becomes a customer. Count it the same way every time, because first order, signed contract, and first payment can differ by weeks or months. ### Loyalty The customer buys again or renews. Measured as repeat rate, renewal rate, or share of wallet. ### Advocacy The customer recommends the brand to others, through referrals, reviews, or sharing. Measured by referral counts or recommendation scores. It feeds awareness for new prospects, which turns the funnel into a loop. ## What it is not - A model of how people buy. It is a reporting structure. Real decisions loop, stall, and restart. - A guarantee that customers pass through every stage in order. People buy on impulse without consideration, and business buyers move in groups. - The sales pipeline in a CRM. Pipeline stages are steps the seller defines, such as meeting held or proposal sent. Funnel stages are meant to describe where the buyer is. ## The steps 1. **Define the customer and the purchase.** Choose the segment, and write down what counts as a purchase. Pick a time window at least as long as a typical buying cycle, or the counts will not line up. 2. **Write stage definitions someone can count.** For each stage, name an observable event and its data source, such as survey recall for awareness or a logged quote request for intent. Mark any stage you cannot measure, and say so in the output. 3. **Count a cohort through the stages.** Follow the same group of people, or the same period with a lag matched to the cycle, and compute the rate from each stage to the next. Dividing this month's purchases by this month's leads in a long sale gives a meaningless rate. 4. **Compare to find the leak.** Set the rates against an earlier period, other segments, and other channels. Look for the stage that falls furthest below the comparison. Break that stage down by source and segment. 5. **Investigate the leak and assign owners.** Talk to people who left at that stage, or review lost deals. Write a hypothesis and the test that would settle it. Give each stage an owner, a target, and a re-measure date with unchanged definitions. ## In practice Web visits had risen for three quarters, leads had risen with them, and closed deals were flat. Marketing called it a sales problem, and sales called it a lead quality problem. Neither had stage definitions that matched. We agreed one set of counts for a single quarter: 20,000 site visits, 600 leads (3 percent of visits), 150 qualified by sales (25 percent of leads), 60 demos (40 percent of qualified), and 12 closed (20 percent of demos). Qualification was the outlier. The prior-year rate on the same definition was 40 percent, and the drop began in the month a gated-content campaign started. That narrowed the argument to one stage and one change. It did not prove the campaign was the cause, since other things changed that month. The next step was to split the qualification rate by lead source. If leads from other sources still qualified at 40 percent, the campaign was pulling in the wrong people, and the lead-count target that rewarded it needed changing. ## What you leave with - Stage definitions, with data sources, that marketing and sales have agreed on. - A table of counts and stage-to-stage rates by segment or channel, for a cohort long enough to have closed. - The stage with the largest gap and a hypothesis about the cause. - An owner for each stage and a date to re-measure. ## Failure points ### Counting the wrong cohort This month's purchases are divided by this month's leads in a sale that takes six months. The rate swings with lead volume, and the team reacts to noise. ### Optimizing the top Lead volume is rewarded, so the team buys more of it. Quality falls, the qualification rate drops, and the cost per customer rises while the dashboard shows growth. ### Definitions that differ by team Marketing counts a form fill as a qualified lead and sales counts a budget conversation. Each team's numbers are right by its own rule, and the rates between stages measure the disagreement. ### Fixing the stage where people leave Prospects leave at the demo stage because the ad promised something the product does not do. Work on the demo cannot fix that. The stage where the loss shows is not always where the cause sits. ### Treating the funnel as behavior Returning visitors, long research phases, and buyers who re-enter are dropped as noise or counted twice. The model then reports a clean straight line that the customers never followed. The failures above are usually [Commercial Acumen](https://usefulconsultingframeworks.com/your-edge/#commercial): the work was not tied to money, for the client and/or for you. ## Origin The funnel grew out of AIDA: attention, interest, desire, action. The formula is commonly credited to the advertising writer E. St. Elmo Lewis. The earliest version I could trace had three steps (attention, interest, conviction) and ran anonymously in Printers' Ink in February 1898, so the credit to Lewis rests on attribution. The funnel image was tied to the formula in William W. Townsend's Bond Salesmanship (1924), according to the usual citation. No one owns the funnel and no trademark applies. Later marketers added stages after the purchase, such as loyalty and advocacy, and relabeled the early ones as awareness, consideration, and intent, so the number and names of stages differ by author. In 2009 David Court, Dave Elzinga, Susan Mulder, and Ole Jørgen Vetvik of McKinsey published a circular model of consumer decisions, based on almost 20,000 consumers in five industries and three continents, as a response to the funnel's linear shape. ## Further reading - [AIDA (marketing)](https://en.wikipedia.org/wiki/AIDA_(marketing)). Wikipedia's history of the formula, including the 1898 Printers' Ink item and the disputed credit to Lewis. Free. Read it before you state an origin to a client. - [McKinsey on the funnel's limits (Court, Elzinga, Mulder, and Vetvik, 2009)](https://www.mckinsey.com/capabilities/growth-marketing-and-sales/our-insights/the-consumer-decision-journey). McKinsey Quarterly, June 2009. The best-known critique of the funnel, based on almost 20,000 consumers. It also says what the funnel still does well. Free to read when checked. - [Purchase funnel](https://en.wikipedia.org/wiki/Marketing_funnel). Wikipedia's overview of the stages, the variants, and the online conversion funnel. Free, and a quick way to see how other sources label the stages. ## Related frameworks - [Net Promoter Score](https://usefulconsultingframeworks.com/frameworks/nps/): One number for recommendation and loyalty. Pick it to track the last stages on a single measure, and the funnel to see where the earlier losses are. - [Brand Equity](https://usefulconsultingframeworks.com/frameworks/brand-equity/): Looks at what the brand means to customers. Pick it when the problem is that people do not consider you, and the funnel to find which stage the loss is at. - [Diffusion of Innovations](https://usefulconsultingframeworks.com/frameworks/diffusion-of-innovations/): Describes which kinds of buyers adopt, and when, across a population. Pick it when the question is who comes next, and the funnel when it is where individual prospects leave. Last updated October 19, 2026.