- AEO
- Answer Engine Optimization - being the source AI assistants quote when they answer, not just a blue link.
- GEO
- Generative Engine Optimization - shaping content so generative engines (ChatGPT, Google AI Overviews, Perplexity) surface and cite it.
- AI Overviews / SGE
- Google's generative answers at the top of the results page; they compress classic organic click-through and reward being cited.
- Zero-click search
- A search answered on the results page itself, so the user never clicks through - now the majority of searches.
- llms.txt / machine-readable surfaces
- Files (llms.txt, AGENTS.md, OpenAPI, an MCP card) that make a site legible to AI agents; still rare enough to be an edge.
- E-E-A-T
- Experience, Expertise, Authoritativeness, Trust - Google's quality framework and the bar AI answers implicitly apply.
- SERP
- Search Engine Results Page - the mix of ads, organic links, snippets, maps and AI answers a query returns.
- Search intent
- Why someone searched - informational, navigational, commercial or transactional; matching it is the core of ranking.
- On-page vs off-page SEO
- On-page = content, structure and HTML you control; off-page = links and mentions others give you.
- Backlink / domain authority
- A link from another site, still the strongest ranking signal; 'authority' scores estimate a domain's link strength.
- Technical SEO
- Crawlability, indexation, site speed, canonicals and structured data - the plumbing that gates everything else.
- Core Web Vitals
- Google's page-experience speed and stability metrics (LCP, INP, CLS) that feed ranking.
- Structured data / schema
- Schema.org markup that lets search and AI engines understand a page and show rich results.
- Featured snippet
- A direct answer box pulled from a page to the top of the results - the classic 'position zero'.
- Topical authority
- Covering a subject comprehensively so engines treat the site as an expert, not a one-page dabbler.
- Pillar page & content cluster
- A hub page plus supporting articles interlinked around one topic - the standard content-architecture pattern.
- Local SEO / map pack / GBP
- Ranking in the map results via a Google Business Profile, reviews and consistent citations - the highest-intent local search.
- PPC
- Pay-per-click advertising - you bid and pay only when someone clicks, the model behind search and most social ads.
- CPC / CPM / CPA
- Cost per click, per thousand impressions, and per acquisition - the three ways ad spend is priced and judged.
- CTR
- Click-through rate - clicks divided by impressions; the core signal for ad and snippet performance.
- ROAS
- Return on ad spend - revenue produced per dollar of ad spend; the headline paid-media metric.
- Quality Score
- Google Ads' rating of ad and landing-page relevance that lowers cost and raises position for relevant advertisers.
- Retargeting / remarketing
- Showing ads to people who already visited or engaged - usually the cheapest paid conversions.
- Lookalike audience
- An ad audience a platform builds to resemble your best existing customers.
- Attribution
- Assigning credit for a conversion across touchpoints - first-touch, last-touch or multi-touch; the perennial reporting argument.
- UTM parameters
- Tags appended to a URL so analytics can tell which campaign, source and medium drove a visit.
- Marketing funnel (TOFU/MOFU/BOFU)
- Top, middle and bottom of funnel - awareness, consideration and decision stages that content and offers are mapped to.
- Inbound marketing
- Earning attention with useful content instead of buying it - the model HubSpot named and built a category on.
- Lead magnet / gated content
- A valuable asset offered in exchange for contact details - the classic lead-capture mechanic.
- ICP & buyer persona
- Ideal Customer Profile (the account) and buyer persona (the person) - who the marketing is actually for.
- CTA
- Call to action - the specific next step a page or email asks the reader to take.
- Deliverability & sender reputation
- Whether email reaches the inbox rather than spam, governed by authentication (SPF/DKIM/DMARC) and engagement history.
- Drip / nurture campaign
- An automated sequence of emails triggered by behavior or time to move a lead toward a decision.
- List segmentation
- Dividing an audience by behavior or attributes so messaging is relevant instead of one-size-fits-all.
- A/B test
- Comparing two versions to a live audience to let data, not opinion, pick the winner.
- CAC / LTV
- Customer acquisition cost vs lifetime value - the unit economics that decide whether growth is profitable. In practice marketers separate blended CAC (all new customers against all spend) from paid CAC, watch MER (total revenue divided by total ad spend) as the blunt whole-business ratio, and judge a channel on contribution margin and payback period rather than on ROAS alone.
- CRO
- Conversion Rate Optimization - turning more of the existing visitors into customers.
- MQL / SQL
- Marketing- vs sales-qualified lead - the handoff definitions that keep marketing and sales honest with each other.
- Engagement rate & reach vs impressions
- Reach = unique people, impressions = total views, engagement rate = interactions per view - the social scorecard.
- Influencer marketing / UGC
- Paying or partnering with creators, and user-generated content, to borrow trust and reach.
- Demand gen vs lead gen
- Creating future demand across an audience vs capturing existing intent as contactable leads - different goals, different metrics.
- ABM (account-based marketing)
- Treating a named set of high-value accounts as the market and marketing to them individually - the B2B enterprise play.
- GTM (go-to-market)
- The plan for how a product reaches and wins its market - channels, messaging, pricing and motion combined.
- First-party data & cookie deprecation
- Data you collect directly (with consent) - increasingly the foundation as third-party cookies and cross-site tracking fade. Google's plan to remove third-party cookies in Chrome was delayed repeatedly and then reversed in favour of a user-choice approach, so the cookie has not vanished; signal loss arrived anyway through Safari and Firefox blocking, ad blockers, iOS ATT and consent banners.
- Multi-touch attribution & attribution models
- Assigning conversion credit across every touchpoint rather than one. First-touch credits discovery, last-touch credits the closer, linear splits evenly, time-decay weights the recent, and data-driven attribution lets the platform model each touch's contribution. All of them are models, not measurements - they only see the touches that were tracked, which is why they systematically over-credit the last clickable channel and under-credit brand and dark social.
- Incrementality testing & geo holdout
- The experiment that answers what attribution cannot: would this conversion have happened anyway? Run as a geo holdout or matched-market test - spend in some markets, go dark in comparable ones, compare. Sizing it needs a minimum detectable effect (MDE): with typical conversion volumes a small test can only resolve a large lift, so a 'flat' result often means the test was underpowered rather than the channel was worthless.
- Marketing mix modeling (MMM)
- A regression on aggregate spend and outcomes over time that needs no user-level tracking, which is why it is back. Its two working ideas are adstock (advertising's effect carries into later periods rather than landing all at once) and the saturation curve (each additional dollar in a channel returns less than the last - diminishing returns), which together tell you where the next dollar should go. Open-source implementations exist from Meta (Robyn) and Google (Meridian).
- Brand vs performance & share of search
- The budget split between building future demand and harvesting present intent - the argument that never ends because brand pays back slowly and performance reports instantly. Share of search (your brand's search volume as a share of the category's) is used as a cheap leading indicator of brand health, alongside brand lift studies. Related trap: view-through conversions and long conversion lag both let a channel claim credit that a holdout would not support.
- AI citation share & brand mentions
- The emerging measure of visibility when there is no click: how often an assistant names your brand in its answer, and how often it cites your page as a source. Assistants answer by retrieval-augmented generation - they search, retrieve a handful of passages, and write from those - so the unit that wins is a clearly-scoped, well-sourced passage that answers the question directly, not a page that ranks. Reported as citation share and mention frequency by prompt, because impressions and clicks no longer exist.
- Entity SEO & the knowledge graph
- Search and AI engines reason about entities - people, organisations, products - not strings. Being a resolved entity (consistent name, an About page, sameAs links to authoritative profiles, schema.org markup) gives a machine something to ground an answer on. Its companion discipline is content pruning: removing or consolidating thin pages so the site's topical signal is not diluted by pages nobody should be retrieving.
- Consent mode, server-side tagging & the conversions API
- The plumbing marketers rebuilt after browser signal loss. Consent mode adjusts what tags send based on the user's consent choice; server-side tagging moves collection from the browser to your own server; conversions APIs and enhanced conversions send hashed first-party identifiers back to ad platforms so conversions can still be matched. Clean rooms (Google Ads Data Hub, Amazon Marketing Cloud) let two parties analyse overlapping audiences without either handing over raw customer records - identity resolution without identity sharing.
- Privacy & messaging law for US marketers
- The rules that govern the list, not the creative. CCPA/CPRA in California plus a growing set of other state privacy laws create opt-out and deletion rights; GDPR still reaches US firms who market to people in the EU. On outbound: CAN-SPAM sets the rules for commercial email, and TCPA requires prior express written consent for marketing SMS and autodialled calls, with statutory damages per message - which is why bought SMS lists are a genuinely dangerous idea.