The marketing dictionary, decoded.

An experiment that splits traffic between two versions — control and variant — so the difference in outcome can be attributed to the one thing you changed.

Most “winning” A/B tests in the wild are noise. If you did not decide the sample size before starting, you did not run a test.

Everything visible on a page or ad before the user scrolls. Inherited from newspaper layout, still the highest-value real estate you own.

The fold moves with every device. Design for the first three seconds, not a fixed pixel height.

Treating a named list of high-value accounts as the market, and building bespoke campaigns for each — the inverse of casting wide and filtering later.

Only worth it when deal sizes are large enough that one closed account pays for the whole programme.

The moment a new user first experiences the core value of the product — the message sent, the report generated, the first order delivered.

Define it as a specific event, not a feeling. Teams that cannot name their activation event cannot improve retention.

The score an auction uses to decide whose ad shows and in what position. Bid is only one input; relevance and expected engagement are the others.

This is why a better ad can outrank a bigger budget — and why creative is a bidding lever, not a design task.

Paying partners a commission for sales they refer, so media cost is incurred only after revenue arrives.

Watch for affiliates bidding on your brand name or dropping last-click cookies — you can end up paying commission on customers you had already won.

An AI system that plans and executes a multi-step task using tools — searching, calling APIs, writing files — rather than answering a single prompt in one shot.

The practical unlock for marketers: research, drafting, QA and reporting chains that run unattended and only escalate on exceptions.

Attention → Interest → Desire → Action. The oldest funnel model in advertising, and still the fastest way to audit whether a piece of creative is doing a job.

Useful as a diagnostic, dangerous as a plan — real buyers loop, pause and re-enter rather than descending tidily.

Presenting a higher reference price first so the price you want to sell at feels reasonable by comparison.

The anchor has to be credible. A fake “was” price that nobody ever paid damages trust faster than it lifts conversion.

The rules used to assign credit for a conversion across the touchpoints that preceded it — first-click, last-click, linear, time-decay, data-driven.

Attribution is a modelling choice, not a measurement. Two honest analysts with different models will report different winners from identical data.

How long after seeing or clicking an ad a conversion still gets credited to it — commonly 1, 7 or 28 days.

Two platforms with different windows will both claim the same sale. Normalise windows before comparing any two channels.

The real-time mechanism deciding which ad a specific person sees. Every impression on Meta, Google and Amazon is priced independently, thousands of times per second.

You never buy an audience — you win access to individual impressions. Costs rise when competitors want the same person, not when the platform “changes the algorithm”.

Dividing a customer base into groups that behave differently enough to deserve different treatment — by value, recency, lifecycle stage, source or intent.

A segment is only real if you would send it a different message. Otherwise it is a chart, not a segment.

The mean revenue per order. One of only three levers on revenue: more customers, more orders each, or more value per order.

Raising AOV raises the CAC you can profitably afford — often cheaper than winning the auction on price.

The point at which someone knows they have a problem but not that your solution exists. Content here teaches; it does not sell.

Judging awareness campaigns on last-click ROAS is the single fastest way to kill future demand.

A reference figure — industry, competitor or historical — used to judge whether your own number is good.

Your own trend line is a better benchmark than an industry average built from businesses with different margins, prices and audiences.

A hard ceiling on what the platform may bid for a single result, used to defend a cost target rather than chase volume.

Set too low it starves delivery entirely. A cap that spends nothing is not discipline, it is an account that stopped working.

The instruction you give the auction: maximise clicks, maximise conversions, hit a target CPA, hit a target ROAS, or bid manually.

Automated bidding needs conversion volume to learn. Below roughly 30–50 conversions a month, a target strategy is guessing expensively.

Acquisition cost calculated across all spend and all new customers, including organic and referral — the number a CFO cares about.

Paid CAC flatters you by taking credit for free customers. Report both, and never mix them in one sentence.

The decision stage: comparison pages, pricing, demos, retargeting, brand search. Demand you are capturing, not creating.

BOFU always looks the most efficient because it harvests intent someone else generated. Scaling it alone hits a ceiling fast.

The share of sessions that end without a second interaction. In GA4 it is simply the inverse of engagement rate.

A high bounce rate on a page that answered the question completely is a success, not a failure. Read it alongside intent.

The commercial premium a name earns beyond the product itself — measurable as pricing power, preference at parity, and lower cost to acquire.

The clearest financial proof of brand equity is a falling CAC while spend stays flat.

The measured change in awareness, recall or consideration caused by a campaign, established by surveying an exposed group against a held-out control.

It is the only credible way to defend upper-funnel budget in a room that only speaks in conversions.

Controlling the context your ads appear in, so they do not run beside content that damages the brand.

Blocklists that are too aggressive strangle reach and defund legitimate publishers. Review exclusion lists as often as you review creative.

Searches containing your company or product name — the cheapest, highest-converting and least incremental traffic in any account.

Brand search volume is a demand thermometer. When it rises after an upper-funnel campaign, that campaign worked.

The loosest keyword match type: your ad can serve on searches the engine judges related in meaning, not just in wording.

Powerful with strong conversion signal and a disciplined negative list. Without both, it is a budget leak.

How many months of margin it takes to earn back what you spent acquiring a customer.

This is the metric that decides how fast you can grow without external funding. Under 12 months is the usual comfort line.

When a channel takes credit for conversions another channel would have delivered anyway — classically, brand search claiming customers who were already coming.

Test it by switching the suspect channel off in one region. Revenue that does not move was never being created.

The share of customers who stop paying in a period. The silent tax on every acquisition budget.

Cutting churn from 5% to 4% monthly extends average lifetime from 20 months to 25 — a 25% LTV increase with zero extra spend.

The share of people who clicked after seeing your ad or link.

A weak CTR is almost always a creative or targeting problem. Raising the budget makes it worse, not better.

Grouping customers by when they joined and tracking each group over time, so you can see whether the business is genuinely improving or merely growing.

Blended averages hide decay. Cohorts are where you discover that this quarter’s customers are worth less than last quarter’s.

A framework where tags adjust their behaviour based on what a visitor consented to, modelling the gaps rather than simply losing the data.

Consent rate is now a performance variable. A poorly designed cookie banner can quietly cost more conversions than a bad creative.

Where a buyer knows solutions exist and is actively comparing them. Content here differentiates rather than educates.

Comparison and alternatives pages are the most under-built, highest-intent assets in most content plans.

What is left from an order after every variable cost of delivering it, including the ad spend that won it.

The metric every other metric answers to. Optimise ROAS long enough and you will eventually optimise CM to zero.

The share of visitors or clicks that complete the action you care about.

Doubling CVR halves CPA without touching the auction — usually the cheapest performance win available.

The period after an ad interaction during which a conversion is still counted — the setting that decides how generous your reporting is.

Match the window to your real buying cycle. A 1-day window on a 6-week B2B decision will make every channel look broken.

Sending conversion events to an ad platform from your server rather than the browser, so they survive ad blockers, cookie loss and iOS restrictions.

Not just a reporting fix — better signal directly improves how well automated bidding can optimise.

Google’s measurable page-experience signals: loading speed, interaction responsiveness and visual stability.

A minor ranking factor but a major conversion factor. Fix them for revenue and take the SEO benefit as change.

What one conversion costs inside a campaign. Distinct from CAC, which is a whole-business number.

CPA is a platform metric and CAC is a P&L metric. Confusing them is how teams report profit that never arrives.

The average price paid per click, set by auction competition and your relevance.

Rising CPCs are a market signal, not a failure. What matters is whether CPA moved with them.

What one captured lead costs in media.

Optimising CPL alone reliably produces cheaper, worse leads. Pair it with the MQL→SQL acceptance rate or it will mislead you.

The cost of one thousand impressions — the price of attention before anyone has done anything.

CPM tells you what the market charges for this audience. If CPM is stable and CPA is climbing, the problem is on your side of the click.

The decay in performance as an audience sees the same asset repeatedly — visible as rising frequency, falling CTR and climbing CPA on unchanged targeting.

The most common cause of “the algorithm broke”. Diagnose with frequency before you touch the budget.

The fully loaded cost of winning one paying customer — media, tools, agency fees and the salaries of the people doing the work.

If your CAC excludes salaries it is not CAC, it is CPA wearing a suit.

The whole sequence of interactions a person has with a brand, from unaware stranger through to repeat advocate.

Map it from the customer’s point of view, not your org chart. The journey does not care which team owns which channel.

Total gross profit a customer generates across the whole relationship. The ceiling on what you can afford to pay to acquire them.

Forecast LTV on a 12-month window rather than “forever”. Infinite-horizon LTV is how unprofitable companies justify unprofitable spending.

Every influential touchpoint you cannot track — WhatsApp groups, podcasts, Slack communities, word of mouth, a screenshot forwarded to a friend.

The honest fix is not better tracking. Ask new customers how they heard about you, and treat that answer as data.

A secure environment where two parties match audience data for measurement without either side seeing the other’s raw records.

How large advertisers still measure overlap and incrementality with walled-garden platforms in a privacy-restricted world.

Scheduling ads to run only in the hours or days that convert, instead of spreading budget evenly across the week.

Check the data before cutting hours — late-night browsing often converts the next morning, and the schedule takes the blame.

Whether your email reaches the inbox at all. Governed by sender reputation, authentication records, engagement history and complaint rate.

Sending to unengaged contacts to hit a volume target is how a healthy list becomes a spam-foldered one.

Reaching people already looking for what you sell — search, marketplaces, comparison sites, retargeting.

Capture is limited by how much demand exists. When you have taken all of it, the only way up is to create more.

Creating want where none existed — the content, social and video work that makes someone realise they have a problem worth solving.

Generation pays off in capture channels. Cut it and search volume for your brand quietly declines a quarter later.

Sessions with no recorded referrer — nominally typed URLs, but in practice a bin for untagged links, apps, PDFs and stripped referrers.

Growing direct traffic is often a tagging failure, not a brand win. Audit your UTMs before you take the credit.

The first real sales conversation, spent diagnosing the buyer’s situation rather than presenting the product.

The strongest predictor of a closed deal is how much the buyer talked. A discovery call that turns into a demo has already gone wrong.

A pre-built sequence of messages released on a schedule or triggered by behaviour — onboarding, nurture, win-back, abandoned cart.

Behaviour-triggered beats time-triggered almost every time. Relevance is a function of timing, not copy length.

Letting the platform assemble ads from a pool of headlines, images and CTAs, then serve the combination most likely to work for each person.

DCO multiplies the assets you give it. Feed it six variations of one idea and it will find the best version of a mediocre idea.

Coverage and attention you did not pay for — press, reviews, organic shares, creator mentions, community recommendations.

The most trusted and least controllable channel. You cannot buy it, but you can be worth talking about.

The share of delivered emails recorded as opened, measured by a tracking pixel.

Apple Mail Privacy Protection pre-fetches images, inflating opens. Treat open rate as directional and optimise on clicks and revenue.

Interactions as a share of the audience that saw the post. The clearest signal of whether content earned attention or merely occupied it.

Check whether the denominator is reach or followers. Agencies quoting the flattering one is a genuine industry habit.

Material that stays useful long after publication — guides, definitions, calculators — and compounds traffic instead of spiking and dying.

Evergreen still needs maintenance. An unrefreshed guide slowly loses rankings to a maintained competitor.

Detecting that a visitor is about to leave — cursor to the tab bar, rapid scroll-up — and triggering a last offer.

It recovers some revenue and annoys everyone else. Cap the frequency and never fire it on a visitor already converting.

The share of sessions that ended on a given page. Unlike bounce rate, it counts visitors who arrived from elsewhere on the site.

A high exit rate on a thank-you page is correct. On a checkout step it is where your revenue is leaking.

Additional revenue from existing customers through upgrades, seats, usage or add-ons.

It carries no acquisition cost, so a rupee of expansion is worth more than a rupee of new business. Most teams under-resource it badly.

How many valid tests a team completes per month. In mature growth orgs it predicts results better than budget does.

Learning rate compounds. Four tests a month beats one perfect test a quarter, even at a lower win rate.

Further training a base model on your own examples so it reliably adopts a specific tone, format or task without being told every time.

Usually the wrong first move. Try better prompting and retrieval first — they solve most brand-voice problems at a fraction of the cost.

Information collected directly from your own audience with consent — purchases, site behaviour, email engagement, survey answers.

As third-party cookies disappear, first-party data stops being a nice-to-have and becomes the targeting infrastructure itself.

Assigning all conversion credit to the first interaction a person had with the brand.

Useful for judging what creates demand, useless for judging what closes it. Run it alongside last-touch and read the gap between them.

Giving a capable version away permanently and charging for depth, scale or collaboration. Acquisition becomes a product function rather than a media buy.

Only works when the free tier creates a habit that hurts to lose. Otherwise it is a discount with extra steps.

The average number of times one person saw your ad in a period.

The first number to check when performance decays. Past roughly 3–4 on a cold audience you are usually paying to annoy people.

A limit on how many times one person may see your ad within a period.

The cheapest defence against creative fatigue and negative brand sentiment. Set it before performance decays, not after.

Deliberately funding awareness, consideration and conversion together, and judging each stage on its own appropriate metric.

Efficiency drives usually cut the top of the funnel first, which is why the following quarter’s conversion campaigns mysteriously get more expensive.

The staged model of how strangers become customers — awareness, consideration, conversion, retention — and the shared language most marketing plans are written in.

Real journeys are messy loops. The funnel is a planning tool; do not mistake the diagram for the customer.

Google’s event-based analytics platform, where everything is an event with parameters rather than a pageview with dimensions.

The event model is only as good as the naming plan behind it. Design the taxonomy first or you will be renaming events for a year.

Material placed behind a form so access is exchanged for contact details.

Gating cuts reach by roughly an order of magnitude. Gate the tool or the benchmark data; leave the thinking open where it can build reputation.

Optimising to be cited inside AI-generated answers — in assistants and AI search results — rather than only ranking in a list of blue links.

Rewards clear structure, explicit definitions, original data and quotable statements. Keyword stuffing does nothing here.

Running a campaign in some regions and withholding it in comparable ones, then reading the revenue difference as true causal impact.

The most practical incrementality test available to a mid-sized brand — no data science team required, just discipline about the control set.

The share of revenue left after the direct cost of delivering the product.

Margin sets your maximum viable CAC. A 70% margin business can outbid a 25% margin business for the same customer, permanently.

A self-reinforcing cycle where the output of one turn becomes the input of the next — users create content, content attracts users, who create more content.

Funnels leak and need refilling; loops compound. Every durable growth story is a loop, not a campaign.

A metric you monitor to make sure a win somewhere is not causing damage elsewhere — refund rate while chasing conversions, unsubscribes while chasing sends.

Set guardrails before the experiment, not after the argument.

When a language model produces fluent, confident output that is factually wrong — invented statistics, fake citations, plausible non-existent features.

The reason AI-assisted marketing needs a verification step, not just a review step. Fluency is not accuracy.

When activity in one channel lifts results in another — a YouTube campaign raising branded search, a retail launch lifting online sales.

Halo is invisible to last-click reporting, which is precisely why brand budgets get cut in efficiency drives.

An email permanently rejected because the address does not exist. Distinct from a soft bounce, which is a temporary failure.

Sustained hard bounces destroy sender reputation. Purge them immediately — a smaller list that lands beats a bigger one that does not.

Systematically trying alternative headlines because the headline carries more of the performance than any other single element.

Test different angles, not different wordings. Ten phrasings of one promise will all perform the same.

Deliberately excluding a random slice of the audience from a campaign so their behaviour becomes the true baseline.

Uncomfortable and correct. A permanent 5% hold-out is the cheapest insurance against believing your own attribution.

The share of viewers still watching after the first three seconds — the single strongest predictor of whether a video ad will scale.

Fix the hook before anything else. No amount of budget rescues an ad nobody watches past the opening frame.

A precise description of the accounts that buy fastest, stay longest and expand most — built from your own closed-won data, not from aspiration.

A sharp ICP is the highest-leverage document in marketing. It makes targeting, messaging and disqualification all easier at once.

The percentage of available auctions you actually appeared in, and why you missed the rest — budget or rank.

The clearest indicator of headroom. Lost impression share to budget means demand exists that you are choosing not to buy.

The number of times your content was served, counting repeat views of the same person.

Served is not seen. Ask for viewable impressions before treating the number as attention.

A platform-built segment of people whose recent behaviour suggests they are actively shopping in a category.

Convenient and crowded — you are bidding against every competitor targeting the same obvious segment.

Earning attention with content people seek out, rather than buying interruption. Slower to start and cheaper to sustain.

Inbound compounds but has a long lag. Fund it from a budget that is not judged monthly, or it will be cancelled before it works.

The conversions that happened because of a campaign and would not have happened otherwise. The only question attribution cannot answer.

Retargeting typically reports the best ROAS and delivers some of the lowest incrementality — you paid to reach people already on their way.

Paying creators for access to the trust they have built with an audience, rather than for raw reach.

Audience fit beats follower count every time. A niche creator with 20k engaged followers routinely outperforms a generalist with a million.

Signals suggesting an account is actively researching a purchase — review-site visits, competitor comparisons, surging topic consumption.

Valuable for timing, weak for targeting. It tells you when to reach an account you had already decided to pursue.

Whether search crawlers can execute your JavaScript to see content that is not in the initial HTML response.

Rendering is queued and unreliable. If a page must rank, ship its content server-side rather than trusting the crawler to run your app.

The idea that customers “hire” a product to make progress in a situation. You compete with every alternative that does the job, including doing nothing.

Reframes competitors usefully: a productivity app competes with a notebook and with apathy, not only with other apps.

Documenting every step, channel and emotion a customer passes through from first contact to advocacy, including the parts you do not control.

The value is in the gaps it exposes — the silent week after purchase, the unanswered question before checkout.

A metric explicitly chosen to judge whether an effort is working — as opposed to the dozens you merely monitor.

If everything is a KPI, nothing is. Three per team is usually the honest maximum.

Several of your own pages competing for the same query, splitting authority and confusing the engine about which to rank.

Usually fixed by consolidating into one stronger page and redirecting the rest — not by writing a fourth article on the topic.

How closely a search must match your keyword before your ad can serve: broad, phrase or exact, each trading reach against control.

Match type is the volume dial and negatives are the brake. Running broad without negatives is driving without one.

The date beyond which a language model has no built-in information, because its training data ends there.

Anything after the cut-off must be supplied in the prompt or retrieved. Ask a model about last week’s ad policy and you get confident fiction.

The single page a campaign sends traffic to, built for one action with the navigation and distractions deliberately stripped out.

Message match is everything. If the headline does not echo the ad that was clicked, you pay for the click and lose the visitor.

A model trained on vast text to predict language, which in practice can draft, summarise, classify, translate and reason across marketing tasks.

Treat it as a fast junior with no memory and no accountability: brilliant at first drafts, never the final reviewer.

Giving all credit to the final interaction before conversion. Still the default in most reporting.

It systematically over-credits brand search and retargeting, and under-credits everything that created the demand in the first place.

Something valuable enough that a stranger will trade contact details for it — a template, calculator, benchmark report or checklist.

The best lead magnets solve a narrow problem completely. Broad ebooks generate lists; specific tools generate buyers.

Staying useful to a lead who is not ready yet, so you are the obvious call when they become ready.

Most leads are not rejections, they are wrong-timing. Nurture is how you stop paying to acquire the same person twice.

Assigning points for fit and behaviour so sales works the leads most likely to close instead of the ones that arrived most recently.

Score fit and intent on separate axes. A perfect-fit account doing nothing needs marketing; a poor-fit account doing everything needs disqualifying.

Treating acquisition, onboarding, retention, expansion and win-back as one coordinated programme rather than separate campaigns.

The stage after purchase is where margin lives, and it is almost always the stage with the smallest budget and the fewest people.

A platform-built audience of people statistically similar to a source list you supply — buyers, subscribers, high-value customers.

Output quality is capped by input quality. A lookalike of “all leads” finds more leads; a lookalike of “top 10% LTV” finds more revenue.

Lifetime value divided by acquisition cost — the clearest single read on whether growth is creating or destroying value.

A very high ratio is not a trophy. It often means you are leaving profitable growth unbought.

Total revenue divided by total marketing spend, ignoring attribution entirely.

Immune to tracking loss and platform self-reporting. When ROAS and MER disagree, MER is telling the truth.

A statistical model that estimates each channel’s contribution to revenue using aggregate historical data, without any user-level tracking.

Privacy-proof and directionally powerful, but it needs years of data and cannot tell you which creative to run tomorrow.

A lead that has met the agreed bar of fit and engagement to be handed to sales.

The definition must be co-signed by sales. An MQL only marketing believes in is a number, not a lead.

The connected set of tools a marketing team runs on — CRM, ESP, CDP, analytics, ad platforms, automation.

Most stacks fail at the joins, not the tools. Audit what data moves between systems before buying another one.

The craft of purchasing attention efficiently — choosing platforms, structuring campaigns, setting bids and managing budget against a target.

As bidding automates, the buyer’s edge moves upstream: audience strategy, creative volume and clean measurement.

How closely a landing page echoes the promise, wording and visual of the ad that brought someone there.

The highest-return, lowest-effort conversion fix in existence. Read your ad, then your page, and cut the mismatch.

A smaller action on the way to the real goal — a size selected, a video finished, a pricing page viewed.

Essential when the main conversion is rare. They give automated bidding and your own tests enough signal to learn from.

The evaluation stage, where a known problem meets competing solutions — case studies, webinars, comparisons, free trials.

The most neglected stage in most funnels, and usually where the largest conversion gains are hiding.

Distributing conversion credit across several touchpoints rather than awarding all of it to the first or last.

Weakened badly by cookie loss and cross-device journeys. Increasingly used alongside MMM and lift tests rather than alone.

Testing several elements and their combinations simultaneously to find which interactions matter, not just which single change wins.

Needs far more traffic than an A/B test. Below serious volume it produces confident nonsense.

Paid placements designed to match the form and feel of the surrounding editorial content.

Disclosure is not optional and does not have to hurt. Audiences forgive advertising; they do not forgive being tricked.

Terms you explicitly block so your ads never serve on them — “free”, “jobs”, “reviews”, competitor names you do not want.

The highest-ROI hour in paid search is reading the search-terms report and adding negatives. Most accounts have never had it done properly.

A loyalty measure from one question: how likely are you to recommend us? Promoters minus detractors gives a score from −100 to +100.

The score is nearly useless; the free-text follow-up is gold. Read the comments, not the number.

Revenue from existing customers this period against last, after expansion, downgrades and churn.

Above 100% the business grows without acquiring anyone. It is the number that most separates good SaaS from great SaaS.

Splitting sessions by whether the visitor has been before — two audiences with completely different intent sharing one report.

Site-wide conversion rate is an average of these two. Segment before you conclude anything about a page.

The one measure that best captures delivered customer value and predicts long-term revenue — nights booked, songs played, messages sent.

Choose one that customers would feel if it fell. Revenue itself makes a poor north star — it lags everything you can actually control.

How the thing you sell is packaged, priced, bundled and guaranteed — the structure around the product rather than the product.

A stronger offer beats better ads almost every time. When performance stalls, look at the offer before the targeting.

An experience where channels share context, so a customer can start on mobile, ask in WhatsApp and finish in store without repeating themselves.

Multichannel means being present everywhere. Omnichannel means remembering. Most brands claiming the second are doing the first.

The guided path from signing up to first real value. The period where most churn is decided and almost no budget is spent.

Define the activation moment precisely, then engineer everything toward reaching it faster. Time-to-value predicts retention better than feature count.

Explicit permission to contact someone. Double opt-in adds a confirmation step that verifies the address and the intent.

Double opt-in shrinks the list and improves almost every metric that follows it, including whether you reach the inbox at all.

People who saw your content without you paying for the impression.

Follower count no longer buys reach. Modern feeds are interest graphs, so each post competes on merit rather than on audience size.

Traffic from unpaid search results — historically the most compounding acquisition channel available.

As AI answers absorb simple queries, organic click volume falls while organic influence persists. Measure assisted conversions, not just sessions.

Channels you fully control — site, blog, email list, community, app. No landlord, no algorithm change.

Rented audiences can be revoked overnight. Converting reach into an owned list is the most under-valued job in social.

Any distribution you buy — search, social, display, video, influencer, retail media. Attention rented rather than earned.

Paid media buys reach, not preference. It accelerates a brand people already want and exposes one they do not.

Google’s fully automated campaign type that serves across every inventory — search, shopping, YouTube, display, Gmail, Maps — from one asset pool and one goal.

You control inputs, not placements. Feed quality, asset variety and exclusions are the whole job.

A composite portrait of a target buyer — goals, constraints, decision criteria and objections.

Build them from interviews, not imagination. A persona with a stock photo and a fictional hobby has never improved a single campaign.

Adapting content, offers or timing to what you know about someone — from a first name in a subject line to a fully reordered product feed.

Personalising the offer beats personalising the greeting. The name in the subject line stopped impressing anyone a decade ago.

How fast revenue moves through the sales pipeline, combining volume, value, win rate and cycle length into one figure.

Shows you four levers at once. Shortening the cycle often beats adding leads, and costs nothing in media.

The place you deliberately occupy in a buyer’s mind relative to alternatives — who it is for, what it replaces, why it is different.

Weak positioning shows up as expensive media. When nobody can say why you are different, every click has to do the persuading.

Everything that happens after the click — speed, message match, clarity, form length, trust signals, checkout.

Teams spend 90% of their attention pre-click and lose most of the money post-click. The cheapest wins are almost always on this side.

How much demand changes when price changes. Inelastic demand barely moves; elastic demand collapses.

Test price on a small segment before assuming you compete on it. Most brands are less elastic than their discounting habits suggest.

Making the product itself the main acquisition, conversion and expansion engine — free trials, self-serve, in-product invitations.

PLG does not remove marketing; it relocates it into the product. Someone still has to design the moment that earns the upgrade.

Buying ad inventory automatically through real-time auctions across thousands of sites and apps, rather than negotiating placements directly.

Demand strict placement exclusions and viewability reporting. Programmatic scale without hygiene funds a lot of fraudulent inventory.

Structuring instructions, context and examples so a model produces reliable output — role, task, constraints, format, worked examples.

The highest-leverage habit is supplying real examples of the output you want. Description tells; demonstration works.

Visitors who match your ICP and show genuine intent — as opposed to sessions that merely raise the traffic chart.

Traffic growth with flat revenue is not a win. Segment by qualified sessions before celebrating a spike.

Interviews, session recordings, surveys and support transcripts — the evidence that explains why the numbers moved.

Analytics tells you where people drop off; ten customer calls tell you why. You need both, and teams usually only fund one.

Google’s 1–10 diagnostic of how relevant your keyword, ad and landing page are to a searcher.

A higher score buys the same position for less. Relevance work is a discount you grant yourself.

What a searcher actually wants — to learn, to compare, to find a specific site, or to buy.

Intent mismatch is the most common reason good content does not rank. Look at what currently ranks; that is the intent the engine has decided on.

The number of distinct people who saw your content at least once — unlike impressions, which count repeats.

Reach and frequency are the two halves of impressions. Read them together or you will mistake repetition for growth.

Systematically encouraging existing customers to bring new ones, usually with a two-sided incentive.

Referral only scales on top of genuine satisfaction. Incentivising referrals for a product people tolerate just accelerates the bad word of mouth.

Serving ads to people who already interacted with you — visited a page, abandoned a cart, watched a video.

Always reports beautiful ROAS and rarely creates much incremental revenue. Cap the spend and test it against a hold-out.

A plot of what share of a cohort is still active over time. The shape matters more than any single point on it.

A curve that flattens means you have a real product with a loyal core. One that keeps falling to zero means you are renting users, not keeping them.

The share of customers still active at the end of a period. The mirror image of churn, and the foundation under LTV.

Retention is compounding acquisition. Every point you hold is a customer you never have to buy twice.

Giving a model your own documents at question time so it answers from your material rather than from memory.

The practical way to build an assistant that knows your brand guidelines, pricing and past campaigns — and cites them.

Revenue attributed to advertising divided by the advertising spend that earned it.

Calculate your break-even ROAS once and pin it above the desk. At 30% margin you need 3.3× just to stand still.

Unifying the systems, data and processes behind marketing, sales and customer success so revenue is managed as one pipeline.

Most “marketing is not delivering” arguments are RevOps problems: two teams reading different definitions from different tools.

Enhanced search listings — ratings, FAQs, prices, images — generated from structured data on your pages.

They raise click-through without raising rank. Structured markup is one of the few SEO tasks with a fast, visible payoff.

Average time from first qualified contact to closed deal. The hidden multiplier on how quickly marketing spend turns into cash.

Marketing shortens cycles by pre-answering objections. Every question a buyer resolves alone is a week off the calendar.

A lead sales has examined and accepted as a genuine opportunity worth pursuing.

The MQL→SQL acceptance rate is the honest scoreboard for lead quality. Below 30%, marketing is generating volume, not demand.

Earning visibility in unpaid search results through technical health, content quality and credible external references.

With AI answers absorbing informational queries, SEO is shifting toward original data, opinion and depth — things a summary cannot replace.

A defined group of customers who share a trait strong enough to justify different messaging, offers or timing.

Recency, frequency and monetary value remain the fastest useful segmentation in commerce — three columns you already have.

Serving ads in a deliberate order, so the second message assumes the first was seen.

Turns repetition into a narrative instead of nagging — the practical alternative to showing one asset until it burns out.

The search engine results page — now a dense mix of AI summaries, ads, shopping units, videos, maps and only then the classic links.

Rank one is worth far less than it was. Read what the page actually looks like before promising traffic from a position.

Collecting and forwarding analytics events from your own server rather than the visitor’s browser.

Recovers signal lost to blockers and browser restrictions, and gives you control over exactly what data leaves your systems.

Your slice of all category advertising, conversation or search visibility relative to competitors.

Holding share of voice above your share of market is the classic condition for growth — and the classic argument for not cutting brand spend.

The industry-wide decline in trackable user data caused by privacy regulation, cookie deprecation and OS-level opt-outs.

The strategic answer is not better tracking. It is first-party data, server-side events and measurement that does not need user-level identity.

Evidence that other people already chose you — reviews, counts, logos, testimonials, visible usage.

Specific beats impressive. One named customer describing a measurable result outperforms a wall of anonymous five-star ratings.

The confidence that a measured difference is real rather than random noise, conventionally set at 95%.

Stopping a test the moment it looks good is called peeking, and it manufactures winners that do not survive contact with reality.

Publishing genuine expertise and point of view to earn credibility ahead of demand — original research, contrarian analysis, hard-won operating detail.

It only works when it risks disagreement. Content that everyone nods at is content nobody remembers.

The unit an AI model reads and writes in — roughly three-quarters of a word in English. Pricing, limits and speed are all measured in tokens.

Useful for budgeting automation. Knowing a 2,000-word brief costs about 2,700 tokens turns AI cost from mystery into a line item.

The widest stage, where people who do not yet know you first encounter the problem you solve.

Judge TOFU on reach, engagement and downstream branded search — never on same-week conversions.

The full revenue opportunity if every possible buyer bought from you. Useful as a ceiling, dangerous as a plan.

The number that matters operationally is the serviceable, reachable slice — the accounts you can actually get in front of this year.

Any moment of contact between a person and the brand — an ad, a search result, a review, an invoice, a support reply.

The touchpoints that shape opinion most are usually the unglamorous ones nobody owns: delivery updates, error states, refund emails.

A small snippet loaded on your site that reports visitor actions back to an ad or analytics platform.

Increasingly blocked, which is why pixels are now paired with server-side events rather than trusted alone.

Where a visit originated — organic, paid, referral, email, social or direct.

Source quality varies enormously. Always read conversion rate by source before deciding a traffic number is good news.

The share of free trials that convert into paying customers.

Usually an onboarding metric wearing a sales costume. Users who reached activation convert; users who never did were never going to.

The event that starts an automated message — a cart abandoned, a plan limit reached, a period of silence.

Behaviour-triggered messages outperform scheduled ones so consistently that the schedule should be your fallback, not your default.

The profit and loss of a single customer or order, isolated from company-wide overhead.

If the unit loses money, scale multiplies the loss. Growth is only a strategy once one customer is profitable.

The share of recipients opting out of a send. The clearest, most honest feedback channel you have.

A spike is a content problem, not a list problem. Read the campaign that caused it before you blame the audience.

Selling a higher tier of what someone already has, or a complementary product alongside it.

Both carry near-zero acquisition cost, which makes expansion revenue the most profitable revenue in the business.

Watching real people attempt a real task on your site and noting exactly where they hesitate or fail.

Five sessions surface most major problems. It is the highest information-per-rupee research available and almost nobody runs it.

Content made by customers or creators rather than the brand — reviews, unboxings, testimonials, native-feeling ad creative.

It outperforms polished production on social because it reads as evidence rather than advertising. Get usage rights in writing before you scale it.

Tags appended to a URL — source, medium, campaign, content, term — that tell your analytics where a visitor came from.

Agree a naming convention and enforce it ruthlessly. “Facebook”, “facebook” and “FB” become three channels and one argument.

A plain statement of the benefit you deliver, for whom, and why it beats the alternative.

If a competitor could put their logo on your value proposition without it becoming false, you have a description, not a proposition.

A number that reliably goes up, feels good in a deck, and changes no decision — impressions, followers, raw pageviews.

The test is simple: if this number doubled, what would you do differently? No answer means it does not belong in the report.

Turning text into numbers that capture meaning, so a system can find related content even when the wording is completely different.

What makes semantic search, deduplication and content clustering work — and why modern search understands intent rather than matching strings.

The share of impressions where the video played to the end.

Read it with hook rate. High hook and low completion means a strong opening attached to a weak middle.

A conversion counted when someone saw an ad without clicking it, then converted within an attribution window.

The most inflatable metric in digital. Check the window length before accepting any view-through number as performance.

How many new users each existing user brings in.

Above 1.0 growth is self-sustaining. Almost nothing sustains K > 1 for long — treat it as a boost, never as the plan.

People with prior contact — site visitors, video viewers, subscribers, past buyers. Cheaper to convert and finite in size.

Warm audiences are a harvest, not a well. Without cold spend refilling them, efficiency decays month over month.

The share of qualified opportunities that end in a closed deal.

A falling win rate usually means lead quality slipped upstream, long before it means the sales team got worse.

A targeted attempt to re-activate lapsed customers before writing them off entirely.

Lapsed customers already know and trust you, so win-back routinely beats cold acquisition on cost — and almost nobody budgets for it.

People recommending you without being asked or paid — the most persuasive and least measurable channel there is.

It is a product outcome that marketing can amplify but cannot manufacture. Give people something specific and repeatable to say.

Chaining tools so a trigger runs a sequence without a human — a form fill that enriches, scores, routes, notifies and schedules by itself.

Automate a process you have already done manually and documented. Automating a broken process just breaks it faster.

Delivering a product as an ongoing subscription rather than a one-time sale — software, hardware, logistics, even furniture.

It changes the marketing job completely: retention and expansion become revenue functions, not support functions.

A machine-readable list of the URLs you want crawled, with hints about when each was last changed.

It speeds up discovery; it does not force indexing. A sitemap full of thin pages advertises your weakest content to the crawler.

Comparing a period against the same period last year, so seasonality cannot disguise the trend.

The only fair comparison for a seasonal business. Month-over-month growth in a peak season is arithmetic, not achievement.

Extracting more value from existing demand through pricing, bundling, merchandising and inventory allocation rather than through more traffic.

Usually the cheapest growth available. A 5% price increase that costs you no volume is a pure margin gain no campaign can match.

Rebuilding the budget from zero each cycle, where every line must be justified again rather than inherited from last year.

Brutal on legacy spend and brutal on long-payback brand work. Pair it with incrementality evidence or it quietly optimises for whatever is easiest to measure.

A search resolved on the results page itself — by an AI summary, featured snippet or knowledge panel — so no website is visited.

Now the majority behaviour for informational queries. Optimise to be the source that gets summarised and named, not just the page that ranks.

Information a customer deliberately gives you — preferences, goals, quiz answers, sizes — rather than behaviour you observed.

The most durable data you can hold: freely given, accurate, and untouched by any privacy change.