Marketing terms, explained plainly.

Most arguments about marketing performance are not disagreements about the numbers. They are two people using the same word for different things — one counting newsletter signups as leads while the other counts quote requests, or one reporting revenue where the other means profit. This is a working glossary of the terms that decide where budget goes: what each one measures, the formula behind it, what a good figure looks like, and the specific way each one flatters you if you let it.

Every term at a glance

The first and last rows have no arithmetic, and they are the two that cause the most disagreement. That is not a coincidence.

What is a lead?

A contact who has given you a way to reach them and a reason to think they might buy. The line worth being strict about is the first half: an email address or phone number handed over deliberately, not an anonymous session in analytics. The second half is where teams disagree, because "a reason to think they might buy" covers both a newsletter signup and a quote request with a budget attached.

Hence the two qualifiers. A marketing qualified lead (MQL) is one marketing believes is worth a sales conversation, based on who they are and how often they have come back. A sales qualified lead (SQL) is one sales has spoken to and confirmed has a real need, some authority and a timeframe. Most B2B pipelines leak at that handover, and almost always because marketing is measured on MQL volume while sales is measured on revenue — so the two optimise for different things and blame each other for the gap.

Worth knowing before you build a follow-up process around it: Gartner's research on the B2B buying journey found buyers spend only about 17% of their total buying time meeting potential suppliers at all, and roughly 5–6% with any single rep when comparing several. Its 2026 sales survey found 67% would prefer to buy with no rep involved. A lead arriving at your form is usually further through the decision than a sales script assumes.

What is a conversion, and a conversion rate?

A conversion is any action you decided counts — a purchase, an enquiry, a signup, a call. The conversion rate is the share of a group who took it: conversions divided by that group, as a percentage. The arithmetic is trivial; the two words doing the work are which group and which action.

That denominator is where most confusion lives. The same 40 enquiries produce completely different percentages depending on whether you divide by all sessions, by sessions that reached a pricing page, or by sessions from one campaign. All are correct. Quoting one without saying which is how two people end up arguing about a number they both measured properly. Sessions and users are not interchangeable either — one person visiting four times is four sessions and one user.

A macro conversion is the thing the business exists to do; a micro conversion is a step towards it, like an add to cart or a pricing page visit. Track both: when the macro number falls, the micro numbers tell you where. Without them, a drop in enquiries is a mystery you can only solve by guessing — which is usually the real reason a site is not converting.

What is a good conversion rate?

Published benchmarks are close to useless here, and it is worth knowing why rather than hunting for a better source. A conversion rate is a ratio of an action to a traffic mix. Two businesses with identical sites and identical offers will report very different rates if one buys high-intent search traffic and the other runs broad social campaigns. The site did not change — the denominator did.

Useful comparisons are internal: this month against last with the same tracking, this landing page against the one it replaced, this source against that one. If you want an external number, take one tied to a specific behaviour. Baymard Institute's average cart abandonment of 70.22%, calculated across 50 studies, tells an e-commerce team where to look in a way that "the average site converts at 2%" never does.

What is click-through rate (CTR)?

Clicks divided by impressions. Five clicks on a hundred impressions is 5% — Google's own definition, identical whether you are measuring an ad, a search result or an email. What differs wildly is what an impression means in each context, which is why CTR benchmarks rarely survive being moved between them.

Google declines to give a universal target, saying only that a good CTR is relative to what you are advertising and on which network, with one floor: a Search Network keyword below 1% is usually the wrong keyword. For organic search the useful benchmark is the shape of the curve by position — First Page Sage's meta-analysis puts position one at 39.8%, position two at 18.7%, position three at 10.2% and position ten at 1.6%. The drop from first to third is far steeper than from third to tenth.

A rising CTR is not automatically good. A headline that over-promises lifts clicks immediately and fills the site with people who leave at the price. The tell is the pair: CTR up and conversions flat means you changed who clicked, not how many bought.

What is CPL and CAC?

Cost per lead is spend divided by leads. Customer acquisition cost is total acquisition spend divided by new customers. The gap between them is your qualification and close rate, and it is usually where the real problem hides: a channel with a cheap CPL and a terrible close rate produces an expensive CAC, and a dashboard reporting only the first will keep recommending you spend more on it.

CAC is wrong in most meetings for one reason — it counts the advertising and leaves out the people. If you stopped trying to acquire customers tomorrow and a cost would disappear, it belongs in CAC: ad spend, agency fees, the tools, and the salaries of everyone doing marketing and sales. Salaries are usually the largest line and the most often omitted. Match the period to your sales cycle too; if deals take two months, this month's customers came from spend two months ago.

There is no universal target. CAC is good when the customer is worth comfortably more than it, over a period you can survive — which makes it a question about lifetime value rather than about the ad account. A low CAC can even be a warning: the cheapest customers to acquire are the ones already looking for you, so heavy spend on branded search and retargeting produces a beautiful number and very little growth.

What is lifetime value (LTV)?

The total profit one customer produces before they stop being one, and therefore the ceiling on what you can afford to spend winning one. The workable version: average order value × purchases per year × years retained × gross margin. A customer spending €200 four times a year for three years at 40% margin is worth €960 in gross profit.

Use margin, not revenue — a revenue-based LTV is inflated by whatever it costs you to deliver, and every budget built on it overspends by exactly that amount. The other common inflation is survivorship: excluding customers who churned in the first month, on the grounds that they were not really customers, produces a figure that describes only the people who stayed.

The convention is an LTV to CAC ratio around 3:1 — roughly three euros of gross margin per euro spent acquiring it. Read the extremes carefully. Above 5:1 usually means underinvestment rather than brilliance; below 2:1 means a small rise in churn or drop in margin flips you into losing money on every sale. Payback period matters as much as the ratio: two businesses with identical numbers are in very different positions if one recovers its cost in two months and the other in fourteen.

What is ROAS, and how is it different from ROI?

Return on ad spend is conversion value divided by ad cost. Spend €2,000, attribute €8,000, and your ROAS is 4x. Google Ads reports the same idea as conversion value per cost, and its Target ROAS bidding works from the inverse — its own documentation gives the example that wanting €5 of sales per €1 spent means a target of 500%.

ROI compares profit with total cost; ROAS compares revenue with advertising cost only. That difference decides everything the moment margin is thin. Break-even ROAS is simply 1 divided by your gross margin: at 40% margin you need 2.5x, at 20% you need 5x. So a campaign reported as a strong 4x is quietly losing money on every order for the second business and comfortably profitable for the first. The ad platform reports both identically, because it does not know your margin.

ROAS also says nothing about what happens after the first order, which is why it should be read next to LTV rather than alone. And chasing a higher ROAS target usually shrinks a campaign — push the target up and the platform bids only where it is most confident, which is the warmest and most already-decided audience. Maximum profit almost always sits at a lower ROAS and higher volume than maximum ROAS does.

What is attribution?

The set of rules deciding which channel gets credit when a sale took six touches. There is no correct answer, only a defensible one — the thing you want to measure, which touch caused the sale, is not observable. You can see that someone saw an ad and later bought; you cannot see the version of events where they did not.

Last click is the default in most tools and it systematically funds the end of the journey: branded search, retargeting and email, the channels that catch people who were already going to buy. Follow that report and you defund the awareness spend that was producing those branded searches, which works for a quarter and then does not. Google Analytics 4 now supports only two models — data-driven, which estimates each touchpoint's contribution by comparing converting and non-converting paths, and paid and organic last click. First click, linear, time decay and position-based were removed in November 2023, so any reporting template built on them is describing a setting that no longer exists.

For a small business, less machinery is better. Below a few hundred conversions a month, data-driven attribution has too little to learn from and multi-touch modelling produces confident-looking noise. Ask people how they found you at the point of enquiry, watch blended cost — total spend over total new customers — and keep the model reports for spotting changes rather than deciding what to cut. Both halves need a correct Google Analytics 4 Setup underneath them.

What is bounce rate?

It changed meaning, and most advice online still describes the old version. Bounce rate used to mean single-page visits, which counted a nine-minute satisfied read as a failure. Google Analytics 4 redefined it as the exact inverse of engagement rate: the percentage of sessions that were not engaged.

A session counts as engaged if it meets any one of three conditions — it lasted longer than 10 seconds, it triggered a key event, or it included at least two page or screen views. That is a genuine improvement, and it means every benchmark from before 2023 is measuring something else entirely. A team celebrating an improved figure after migrating to GA4 is usually just seeing the definition change.

Whether a bounce is bad depends on the page. On a reference page that answers a question in a paragraph, it is a satisfied visitor. On a landing page you are paying to send traffic to, every bounce is money spent on nothing. Message mismatch and slow rendering cause most paid bounces, and people who leave before the page renders land in the bounce bucket while looking like an audience problem — checking Core Web Vitals on the worst pages separates the two in an afternoon.

What is churn rate?

Customers lost in a period divided by customers you had at the start of it. Keep new customers acquired during the period out of the denominator — including them dilutes the rate and makes fast growth look like good retention. Churn compounds, which is the part that surprises people: 5% monthly churn is not 60% a year but about 46%, because you lose 5% of a shrinking base each month.

Report customer churn and revenue churn together, because they diverge whenever your customers are different sizes. Higher customer churn than revenue churn means you are losing small accounts; the reverse means you are losing your largest ones, which is an emergency dressed up as a rounding error. Recurly's subscription research puts a good annual churn rate at roughly 3–5%, with a median near 3.22% for software — but stresses comparing against a similar model and similar revenue per customer rather than a cross-industry average.

The cheapest churn to fix is involuntary churn — customers lost to failed payments rather than to a decision. Recurly puts the median annual involuntary rate at about 1.25%, a meaningful slice of a 3–5% total, lost to nobody's intent. Card updaters, retry logic spread over days and dunning emails that arrive before access is cut recover much of it, and none of that requires product work. Most teams start with the product; the faster win is usually in billing.

Which of these actually matter for you?

Fewer than the list suggests. For most small and mid-sized businesses, four numbers watched together will out-perform a dashboard of twenty: qualified leads per month, what each one cost, what share became customers, and what a customer is worth. Those four can distinguish a traffic problem from a targeting problem from a website problem from a sales problem, which is a distinction no single metric can make.

The rest are diagnostics — you reach for CTR, bounce rate or attribution when one of the four moves and you need to know why. And retention belongs in the set regardless of what you sell: Harvard Business Review's summary of Frederick Reichheld's Bain research puts a 5% increase in retention at a 25–95% increase in profits, and acquiring a new customer at five to 25 times the cost of keeping one. A team fighting a rising acquisition cost is very often looking at the wrong end of a retention problem.

Sources

The definitions and figures above come from these sources, checked July 2026. The glossary table, the thresholds and the interpretation are our own.

• Gartner — The B2B Buying Journey ↗

  B2B buyers spend roughly 17% of the buying journey meeting potential suppliers, and about 5–6% with any single rep when comparing several. Gartner's March 2026 sales survey found 67% would prefer a rep-free buying experience.

• Baymard Institute — Cart abandonment rate statistics ↗

  Average documented cart abandonment of 70.22%, calculated across 50 separate studies.

• Google Ads Help — Clickthrough rate (CTR): Definition ↗

  CTR as clicks divided by impressions, the note that a good CTR is relative to network and offer, and the guidance to replace Search Network keywords below 1%.

• First Page Sage — Google click-through rates by ranking position ↗

  Meta-analysis of published CTR research: 39.8% at position one, 18.7% at two, 10.2% at three and 1.6% at ten.

• Google Ads Help — About Target ROAS bidding ↗

  Target ROAS as the average conversion value wanted per unit of ad spend, with the worked example of a 500% target returning five times the spend in sales.

• Analytics Help — [GA4] Get started with attribution ↗

  The two remaining GA4 models — data-driven and paid and organic last click — and the removal of first click, linear, time decay and position-based in November 2023.

• Analytics Help — [GA4] Engagement rate and bounce rate ↗

  Bounce rate as the inverse of engagement rate, and an engaged session as one lasting over 10 seconds, or with a key event, or with at least two page views.

• Recurly Research — Churn rate benchmarks by industry ↗

  A good annual churn rate of roughly 3–5% across subscription verticals, a median near 3.22% for software, and a median annual involuntary churn of about 1.25%.

• Harvard Business Review — The Value of Keeping the Right Customers ↗

  Frederick Reichheld's Bain research: a 5% increase in retention raises profits by 25% to 95%; acquisition costs five to 25 times more than retention.