Glossary

Every term the product and this help center use, in one alphabetical list. Each entry is a sentence or two and a link to the page that owns the detail — the glossary is where you find out which thing a word means, not the full table. Italicised words inside a definition have an entry of their own.

Two things to know before you use it:

• Where an entry disagrees with a reference page, the reference page is right. Those pages are written against the code; this one is written against them.
• The same word means different things in different parts of the product. That is not sloppiness in the docs — it is in the product, and it costs real money when someone reconciles a payout against the column whose header matches the word they were looking for. Those names are indexed below.

Ad-tech abbreviations — CTR, eCPM, I2C, CCR, CPM, CPC, DSP, SSP, DMP, VPAID, VAST — are industry-standard and stay in English in all 15 interface languages. They are defined in the list below like any other term.

Names that mean two things

Each of these is one word carrying two or more unrelated meanings. The full entry is in the A–Z below; the link goes to the page that owns the detail.

Name	The senses it carries
Revenue	Net as the publisher sees it — the money column of the Publisher column family. Equal to Net unless the network adjusts the publisher’s figures. Metrics
Charges	Gross as the advertiser sees it — the money column of the Advertiser column family. Equal to Gross unless the network adjusts the advertiser’s figures. Metrics
eCPM	Spelled out (eCPM Net, eCPM Gross, eCPM Profit) it is the network’s own; a bare eCPM header means you are in the Publisher or Advertiser column family. Metrics
Requests	Impressions plus unfilled requests on the General report; a capping kind; and four separate request counters on the Advanced report. Metrics
Fill rate	Four different formulas depending on the report — they are not comparable. Metrics
Advanced report	One report type, HADOOP, reachable by four names — three legacy spellings still resolve to it. Analytics reports via API
Capping type	An 18-value enum of kind-fixed-to-entity-level; and an unrelated 4-value enum of the same name in the REST layer. Capping matrix
Blocked	A traffic-quality verdict that enforces nothing; a creative-scan status that drops the bid; and five targeting rule types that hard-exclude an entity. Statuses
Paused	A status glyph meaning a limit or cap is holding delivery; and a serving-state pill meaning the wiring exists but nothing on it is switched on. Statuses
Rules	The category/label matching engine; targeting rules; and price-adjustment rules — three unrelated systems. Targeting dimensions
Priority	A campaign’s six-band priority; and the delivery tier a banner competes in inside a placement. Entities over REST
Custom parameter	A tag parameter, a targeting dimension and an analytics dimension — the same values seen from three places. Targeting dimensions
Channel	Likewise: a tag parameter, a targeting dimension and an analytics dimension. Invocation codes
Action	A conversion goal; a capping and limit kind; and a grid bulk action. Capping matrix
Template	A rich-media creative skeleton; and an admin email/SMS notification template. Banner types
Mail	The Email tag type is labelled Email on its tab, Mail in the preview picker and the downloaded file, and EMAIL in the API. Invocation codes
Universal JS / Mount JS	Their code and API names are AJAX and JS_MOUNT — grep for the label and you find nothing. Invocation codes
Native	A placement type, two creative types, and the Native API — each with its own rules. Placement types

---

A

Action — a conversion event you track (purchase, signup, install), defined per campaign; also called a goal in the API. Two unrelated things share the word: Action capping and Action limits cap conversions of that kind, and a grid bulk action is a toolbar operation on ticked rows. See Conversion actions.

Action limit — a cap on how many conversions of a given action an entity may accumulate before serving stops (used for CPA auto-deactivation). Separate from delivery limits and available on five entity types. See Action limits.

Ad request — a single call from a page, app or player asking the ad server for an ad. Not the same as the Requests metric, which counts requests and impressions.

Ad-serving domain — the host a placement’s tags call. Chosen per placement on the tag Settings tab; with none configured on the network the two JavaScript tag types disappear entirely. Creatives are served from the CDN attached to the same domain. See Invocation code types.

Ad tag — everyday name for an invocation code.

Ad unit — a standard creative size (300×250, 728×90, …), or Custom for a size you type. Required for Local File, External File, and HTML Code on Site and Video placements; never for Native, and shown but never required for OpenRTB. See Creative & ad specs.

Admin-only permission — a permission that exists but is hidden from the role permission matrix, so a network manager looking at Roles cannot grant it to anyone. Every analytics report type except General is in this set, as is the Advertiser column family. See Analytics dimensions.

Advanced report — the report recomputed from the full ad-request logs, HADOOP in the API and Advanced on screen. Three legacy names still resolve to it — HADOOP_ADREQUESTS, BROWSER_OS and HADOOP_SECURITY — so a bookmarked “Security report” URL runs the Advanced report, not a second report type. It has no column families and its own column arithmetic. See Analytics metrics.

Advertiser — the top demand-side entity. Owns campaigns, which own banners. Hierarchy: Advertiser → Campaign → Banner. Also the name of one column family.

AP Profit — Charges − Revenue: the cross-family margin, the two figures each side of the business actually sees. AP is Advertiser/Publisher, not “accounting period”. Distinct from plain Profit, which stays inside the network’s own columns. See Analytics metrics.

API key — a 32-character hex value you mint on your own preferences page and send in the X-Api-Key header (or as a Bearer token) on any REST call; the call runs with your own permissions, and your role needs API Key Access. The analytics report also still takes it as an apiKey query parameter. An alternative to hash authentication. See REST API overview.

Approval — the moderation state of a banner or a site: banners have four states (New, Pending, Approved, Rejected), sites three (no New). The two enums number the shared names differently, so never compare their raw ids. An absent badge means Approved or New. See Statuses & indicators.

Attribution type — how a conversion was credited: Post-click, Post-view, Post-click (expired), Post-view (expired), System, or Unattributed. The two (expired) buckets are zero-revenue diagnostics excluded from every total. See Analytics dimensions.

B

Banner — a creative of any format. The leaf of the demand tree. There are 12 banner types, each gated by its own permission, and which ones a placement offers depends on its placement type. See Banner types.

Beacon — a URL the caller must fire for an event to count. Native API responses carry an impression beacon; an ad you render without firing it is an ad that was never counted and never billed. See Native API.

Bid floor — the minimum acceptable price in an RTB auction. On the outgoing bid request it is the first non-zero value along a six-link chain (banner → placement → site floors, then the banner price, the campaign price, the site CPM threshold), evaluated per country. See OpenRTB integration.

Blocked — three unrelated things:

1. a traffic-quality level. It is a verdict, not an enforcement — nothing in the ad-serving stack reads the score, so a placement marked Blocked keeps serving (Statuses & indicators);
2. a creative-scan status, which is enforced — a BLOCKED creative is dropped from the auction and the runner-up wins (Creative policy protection);
3. the five BLOCKED_* targeting rule types, hard exclusions created by blocking actions rather than picked from the add-rule list (Targeting dimensions).

C

Campaign — the container for banners under an advertiser. Carries pricing, budget, dates, priority and campaign-level targeting and capping.

Capping — limiting how often one user sees or interacts with ads over a period. See the capping matrix.

Capping subject — what counts as “a user” for one cap row: User (cookie-based), User (server-side storage), User (cookie-based + server-side storage), IP, or L.E.S. IP (the last external source IP, seen before proxies). Set per row, not per entity. See Capping matrix.

Capping type — two different enums share this name:

• the product one has 18 values — a kind of cap fixed to one entity level, so Frequency appears four times. Six of them cannot be created by anyone: the three Request values (no role can hold the permission) and the three DSP ones (refused on every path, including REST). For which of the rest each level offers, see the capping matrix;
• a separate 4-value enum in the REST layer (FREQUENCY, CLICK, ACTION, NO_CLICK) names only the kind. Note the singular spellings — they are not the 18-value enum’s names.

Category — a predefined content grouping (Automotive, Finance, …) attached to demand and supply entities and used by Rules for automatic matching. Unrelated to risk category, which names a fraud pattern.

CCR — click-to-conversion rate: conversions ÷ clicks. See Analytics metrics.

Channel — a free-form traffic-source identifier a publisher passes in the ad tag. The same value surfaces three ways: as a default tag parameter, as a targeting dimension, and as an analytics dimension. It is also the only targeting dimension where the MATCH / NOT_MATCH rule verbs do anything.

Charges — Gross as the advertiser sees it: the money column of the Advertiser column family, and the advertiser’s spend. It equals Gross unless the network adjusts the figures shown to the advertiser; then Charges is the adjusted figure and Gross stays the network’s own. See Analytics metrics.

Clearing price — what the winner actually pays, and what the price macros carry. Under second-price clearing it is min( max(deal floor + $0.01, second-highest bid + $0.01), winning bid ) — not the raw price the winner sent. See OpenRTB integration.

Clipboard — the buffer that holds one homogeneous set of entities so you can paste a copy, move, link or unlink them elsewhere. Shows as a floating bar until cleared. See UI navigation.

Column family — which audience’s copy of a metric you are looking at: Basic (the network’s own), Publisher, or Advertiser. Each is a separate permission, and the General report renders each counting metric once per enabled family. The money column is the only one renamed between families, and those really are different columns. See Analytics metrics.

Command palette — the ⌘K / Ctrl+K overlay for jumping to any entity or page by name.

Content replacement — a placement feature that swaps parts of the served markup based on configured match/replace pairs.

Conversion window — how long after a view or a click a conversion still counts (post-view and post-click windows, set per action); also called the attribution window. A conversion that matches outside its window is counted in an expired diagnostic column with zero revenue and triggers nothing.

CPA — cost per acquisition: a campaign pricing type billed per conversion.

CPC — cost per click.

CPM — cost per mille (per thousand impressions).

Creative — synonym for banner. Both words are used throughout the product; the API says banner.

Creative scan — automatic malware/policy scanning of programmatic creatives from the DSP partners a network chooses; a blocked creative never wins the auction. See Creative policy protection.

CTA — call to action: the button text on a native creative (ctaText in the API, $$NATIVE_CTA_TEXT$$ in an adapter template).

CTR — click-through rate: clicks ÷ impressions. On the Advanced report, CTR and Effective CTR have different denominators and neither is wrong. See Analytics metrics.

Custom parameter — a key=value pair carried on the ad request. It reaches you three ways and they are the same values: as a default tag parameter set on the placement, as the Custom Parameter targeting dimension, and as an analytics dimension you can break rows or columns down by. In creatives it is read with $$CUSTOM_PARAM(key)$$.

D

Daisy chaining — sequential banner fallback inside one placement: if the first banner can’t serve, the next in the chain is tried. Also called the waterfall. See Daisy chaining & passback.

Default attribution type — the attribution type an action falls back to when a conversion reaches it with none declared. Unattributed unless set; a fallback, never an override. See Conversion actions.

Default parameters — the values baked into every tag a placement hands out: channel, custom parameters, partner sub-values, a third-party click URL and seven pass-through signals (GDPR Applies, IAB TCF Consent String, App Bundle, Advertising ID (IFA), User ID, Transaction ID, Ad-Loaded Handler). They are set once on the tag Settings tab and save on change. See Invocation code types.

Demand — the advertiser side of the network: advertisers, campaigns, banners.

Demo account — a time-limited trial account: users on the role Epom configures as the demo role get a welcome overlay, a guided tour and read-only example entities, until the trial days run out. See Demo accounts.

DMP — data-management platform. Also a targeting dimension that drills through a provider’s segment tree, and an import/export API surface.

DSP — demand-side platform: the buyer Epom asks to bid. See Ad server vs DSP.

DSP partner — a configured OpenRTB integration your network requests bids from. One bid request is built per eligible banner, not per partner, so two banners on one partner produce two requests. See DSP partners.

E

eCPM — effective cost per mille: money × 1000 ÷ impressions. Which money, and which impressions, depends on the column: the network’s own are spelled out as eCPM Net, eCPM Gross and eCPM Profit, while a bare eCPM header means you are in the Publisher family (Revenue-based) or the Advertiser family (Charges-based). The Header Bidding report’s eCPM divides by fills rather than impressions and is not comparable with any of them. See Analytics metrics and Revenue metrics.

Eligible — on the Advanced report, an event that survived the invalid-traffic filters: eligible impressions, eligible clicks and eligible conversions are each the raw count minus what was rejected. In ad selection, an eligible banner is one that passed targeting and capping and can therefore compete. See Analytics metrics.

Epom support — Epom’s support team, or your account or success manager at Epom: whoever handles what only Epom can change for your network — server-side settings, a new ad-serving domain, API quotas, billing plans, modules switched on per network. Pages say “ask Epom support or your account manager” for these. Not the same as your network manager, the person who runs your ad network’s account — ask them for a card, tab or permission you do not see — or your network administrator, who edits the roles and the admin pages. Nor is it an administrator, which is any user on an admin-based role: a role that reaches the Admin area. Your network manager is often an administrator too, but the words answer different questions — who to ask, versus which pages a role reaches. See Contact Epom support.

Extended filter — the multi-section grid panel that narrows rows by name/ID, status, traffic activity, rules, labels, categories and dates, one section per level. The grid’s basic filter row (Group by, Filter by with a Name box, My objects) sits beside it. See UI navigation.

F

Fill rate — the share of requests that returned an ad, with four different formulas depending on where you read it: impressions ÷ Requests on the General report; eligible impressions ÷ direct requests, and eligible impressions ÷ deliveries, as two separate columns on the Advanced report; and fills ÷ rounds on Header Bidding. Do not compare them. See Analytics metrics.

Fill source — the Header Bidding report’s own group-by dimension: who filled the slot. It is offered by that report only, not by General or Advanced.

Flat — a per-cap-row option that spreads a cap’s allowance evenly across its window rather than letting it be consumed at once. The REST API calls the same field evenDistribution on the typed capping endpoints and flat on the generic ones, and neither accepts the other’s name. See Frequency capping.

Frequency capping — the classic form of capping: at most N impressions per user per period. See Frequency capping.

G

General report — the default report, read from pre-aggregated statistics. It is the one report type with no enum value: omitting statisticType is how you ask for it. It is also the only report type a network manager can grant — every other type’s permission is admin-only. See Analytics dimensions.

Gross — the network’s advertiser-side money column; the network’s advertiser billing totals it. An advertiser sees the same money as Charges. See Analytics metrics.

Group by — the dimensions a report spreads down its rows. Distinct from Split columns by, which spreads dimensions across the columns and is a multi-select. See Analytics dimensions.

H

Hash authentication — the signed-request scheme, one of the three REST authentication methods (with the API key and the session cookie): username, timestamp (Unix milliseconds) and hash = MD5( MD5(password) + timestamp ). A signature is valid only near its timestamp, so a long-running script must recompute the timestamp and hash per request. See REST API overview.

Header bidding — running an auction for a slot before the ad call. Epom is on both sides of it: it answers inbound bid requests at /hb/bid, and it has a Header Bidding report of its own whose metrics (Fills, Unfilled, its own Fill Rate and eCPM) share no arithmetic with the General report. See Header bidding.

HTML5 ZIP — a zipped HTML5 creative uploaded as a template file property. The only server-side rule is that the filename ends in .zip, lower case — the archive’s contents and size are not checked. See Creative & ad specs.

I

I2C — impression-to-conversion: conversions ÷ impressions. See Analytics metrics.

IAB TCF consent string — the encoded consent record from a CMP, passed in on the tag and read with $$GDPR_CONSENT_IAB$$. Epom uses it to gate its own serving; it is not forwarded in the outgoing bid request. Epom’s own GVL vendor id is 849. See Macros.

IFA — the mobile advertising ID, Android AAID or iOS IDFA, whichever the request carries. Passed in as ifa and read with $$IFA$$; the older per-platform macros are deprecated.

Include / Exclude — the base choice on every targeting rule: Include serves only on matching traffic, Exclude never does. On the three yes/no dimensions the choice is the whole rule — there is no value to pick. See Targeting dimensions.

In-stream / outstream / overlay — the three shapes a Video placement can take: in-stream runs inside the content video, outstream in a player the ad itself creates on a page, and an overlay sits on top of playing content.

Invocation code — the snippet generated per placement that a page, app, email or player uses to request ads. Seven types exist; the VAST tags tab is an eighth entry in the strip but is not a code type. See Invocation code types.

L

Label — a free-form tag you attach to entities for filtering, bulk operations and rule matching. Unlike categories, labels are user-defined.

Limit — a total-volume ceiling per entity with a reset period, as opposed to per-user capping. Six metrics — impressions, clicks, conversions, opportunities, budget and requests — across four periods (daily, weekly, monthly, total). A limit of 0 means “no cap”, not “stop”; to halt delivery, deactivate the entity. See Limits.

Linking — connecting banners to placements (many-to-many). A banner serves only on placements it is linked to — or where Rules match. See Hierarchy & linking.

Log report — Impressions Log, Clicks Log and Conversions Log: event lists, not aggregates, so none of the aggregate metric arithmetic applies to them. They have no Group by, only filters. A fourth, Actions Log, exists and is served by the API but never appears in the report-type picker. See Analytics dimensions.

M

Macro — a $$NAME$$ token in creative code, URLs or tags that the server replaces with a live value at serve time. A macro it recognises but cannot resolve becomes empty; one it does not recognise is left exactly as written. See Macros.

Mail — the Email invocation-code type. Its tab says Email, the preview picker and the downloaded file say Mail, and the API value is EMAIL. One type, three spellings.

Markup — what the ad server returns for a request: HTML for a display creative, VAST XML for video, JSON for native. Macros are resolved while it is being built.

Masked — how a shared entity appears when the owner hides its details. In trees and demand grids the name is replaced and the row renders in italics; you cannot open or edit it and bulk actions skip it. In analytics the row renders as TYPE_id — SITE_45, CAMPAIGN_123. The supply grids do not draw the masked style at all, so a read-only shared site looks ordinary there until an action skips it. See Sharing & permission model.

Minimal revenue — a per-row floor under the publisher’s earnings. Both Net and Revenue are read as GREATEST(column, minimal revenue), and Minimal Net is a column showing the floor by itself. See Analytics metrics.

Mount JS — an invocation-code type: a bare function you call with your own container element. Its code and API name is JS_MOUNT, so searching for the label finds nothing. See Invocation code types.

N

Native — three related things gated differently. The Native placement type (id 9); the Local Native and External Native creative types, which a server-side filter allows on native placements only; and the Native API (/ads-api-native), which returns structured creative fields rather than HTML. See Native API.

Native asset — one of the three files a native creative can carry (icon, image, video), each with its own declared size. The declared size is what the placement’s template lays out; the file’s real dimensions are checked against it. See Creative & ad specs.

Net — the network’s publisher-side money column, floored at minimal revenue; publisher payments sum Net. A publisher sees the same figure as Revenue. See Analytics metrics.

Network administrator — whoever holds the admin console of your network: roles and permissions, admin pages, DSP partners, troubleshooting logs. A publisher or advertiser reaches them through their network manager. Platform-side settings are not theirs to change — see Epom support. See Manage users, roles and permissions.

Network Manager — the operational role for day-to-day entity and campaign management. What it sees depends on the permissions your network administrator grants it. See Network manager. In “ask your network manager” it means the person who runs your network’s account, whatever their role — see Epom support.

Network parameter — a named per-partner onboarding value on a DSP partner or an OpenRTB banner, drawn from a set of 135. Despite appearing all over the bid-request documentation, these are not OpenRTB spec fields, and none of them is a fee or revenue share. See OpenRTB integration.

Admin-based role — a role (the stock AdServer Supervisor and Default Admin are examples; networks rename and add them) that reaches the admin area and whichever pages its Admin UI checkboxes allow, but whose permission matrix carries only the user- and ad-server-management groups. It is not the top role, which Epom keeps for its own staff. See Users & roles.

O

OpenRTB — the real-time bidding protocol. Epom speaks it in both directions: outbound to DSP partners (2.5 by default) and inbound as a bidder at /hb/bid (2.6). See OpenRTB integration.

Opportunity — a moment the ad server selected an entity to serve, counted before the impression is confirmed. Capped by the Opportunities limit and reported as its own Advanced-report column. A /hb/bid call consumes one even though it is a bid, not an impression.

Optimization — automatic banner weighting inside a placement: Manual weights, eCPM optimization, or eCPM with burn-up (a learning period before optimization starts). See Optimization.

P

Passback — the fallback code served when no linked banner can fill the request; set per placement. See Daisy chaining & passback.

Paused — two different signals. As a status glyph (amber gauge) it means a limit or cap is currently holding delivery. As a serving-state pill — Paused · N — it means the wiring exists but nothing on it is switched on. An entity that is both switched off and unlinked reads Not serving, not Paused. See Statuses & indicators.

Piggyback — a third-party tracking pixel attached to a campaign, placement or site that fires alongside conversions. See Piggybacks.

Placement — the ad slot on the supply side; the entity ad tags are generated for. There are 8 placement types, with ids 1–7 and 9 — there is no id 8. The type is chosen at creation and cannot be changed. See Placement types.

Placement key — the 32-character hex identifier a tag is fetched by, stable for the life of the placement. It identifies inventory; it is not a secret.

PMP — Private Marketplace: a deal giving specific programmatic demand preferential access at negotiated terms. Carried in the outgoing bid request as imp.pmp, with a deal id, a floor, a buyer seat and an auction type.

Post-click / Post-view — the two attribution arms of every conversion metric. The money arms are not symmetric across column families: Publisher gets the Net side only, Advertiser the Gross side only, and only Basic carries both.

Presets — reusable named groups for targeting: domain groups, IP-range groups and app-package groups. A preset rule references one group by id rather than carrying values inline. See Presets.

Pretargeting — supply-side targeting. Same 38 dimensions as demand-side targeting, the same card, a different name and a separate set of permissions — the card is labelled Pre-targeting on sites, zones and placements. See Pretargeting & blocking.

Price adjustment rule — a conditional rule that multiplies or sets the gross price when traffic matches a condition; cascades advertiser → campaign → banner, so a banner-level rule is applied last and effectively wins. See Price adjustment rules.

Pricing type — how an entity is billed: CPM, CPC or CPA on campaigns; cost-per-day and cost-per-month are placement- and site-side only. See Pricing.

Priority — two different things. A campaign carries one of six priority bands — PREMIUM, STANDARD_HIGH, STANDARD_MEDIUM, STANDARD_LOW, NON_GUARANTEED, HOUSE_ADS. Separately, a banner competes in a delivery tier inside a placement, and higher tiers are tried before lower ones (the waterfall).

Profit — Gross − Net, both from the network’s own column family, and computed from the already-rounded values. Distinct from AP Profit, which crosses families. The Advanced report clamps Profit at zero; the General report does not.

R

Rank — a popularity position from the device-detection database, where lower is better: Browser rank, Hardware rank and Platform rank rank the browser version, the device model and the OS respectively. Available as targeting dimensions (default relation <=, which keeps the top ones) and as analytics dimensions.

Requests — the General report’s Requests column is impressions plus requests that returned no ad, which is why Fill Rate divides by it and can never exceed 100%. The unfilled half is counted separately from impressions and added in. The Advanced report has no column of this name at all, only Direct, Repeated, Filled, Returned and Wasted requests. Request capping is a third, unrelated use of the word — and one no role can reach. See Analytics metrics.

REST API — the programmatic surface. Every path starts with /rest-api/ and ends in .do; parameters are form- or query-encoded, not JSON bodies. A refusal carries a 4xx status (404 for a missing entity, 403 for a missing grant, 400 for a bad value) with a "success": false body; only a partially applied batch still answers 200 with "success": false. See REST API overview.

Revenue — Net as the publisher sees it: the money column of the Publisher column family, and what the publisher earns. It equals Net unless the network adjusts the figures shown to the publisher; then Revenue is the adjusted figure and Net stays the network’s own. The referral commission is computed from Revenue. See Analytics metrics.

Revenue share — the split of earnings between the network and a publisher, configurable per site, zone or placement. This is where the network’s margin on programmatic spend is taken — nothing is deducted from the DSP settlement price.

Risk level / risk category / risk score — a risk score is the numeric fraud probability on a row; a risk level is the band it falls in; a risk category names the fraud pattern (25 of them). Level and category are dimensions you can group and filter by, score is a metric. See Invalid traffic & risk scoring.

Roll-up — the summary totals row of a report. It is computed by its own query over the whole result set, not by adding up the displayed rows, and it recomputes every ratio from summed numerators and denominators — so it will not always equal the sum of a rounded column.

RTB — real-time bidding; see OpenRTB. Also the name of an account tab and of a placement-editor card.

Rule view — a grid display mode that swaps regular columns for rule columns (rules on/off, include/exclude groups, inherited parent rules). See UI navigation.

Rules — three unrelated systems share this word:

1. Rules proper — the category/label engine that matches banners to placements without manual linking (Rules & defaults);
2. targeting rules — the Include/Exclude conditions on the Targeting and Pre-targeting cards (Targeting dimensions);
3. price adjustment rules — see Price adjustment rule.

The Rules active status glyph and the serving-state Rules only pill both mean the first one.

Run-as — an administrator or network manager opening the platform as another user; the user menu shows Exit user while impersonation is active.

S

S2S (server-to-server) — conversion tracking via postback calls between servers instead of browser pixels. Also the name for calling the ad API from your own backend, where clientIp and clientUa must be supplied or targeting cannot work. See S2S and pixels.

Saved view — a named snapshot of a grid’s filters, sort and page size. Stored per user, per grid, in your browser — it does not follow you to another machine, and saving under an existing name overwrites it silently.

Seat — the buyer identity on an OpenRTB bid (seatbid[].seat) and on a PMP deal (wseat). Epom accepts exactly one seatbid containing exactly one bid; anything else is discarded whole.

Second price — an auction where the winner pays just above the runner-up rather than their own bid. Note that a DSP partner’s Second price only setting does not change the clearing rule — it withdraws the banner when it would be the only bidder. See First vs second price.

Self-serve — the mode in which advertisers sign up and run their own campaigns against prototype creatives and CPM floor rates you configure, rather than a network manager building everything for them.

Serving state — whether an entity can actually deliver right now, shown as the header pill on banner and placement editors: Not serving, Rules only, Paused · N, Serving · N, or Serving · A/T when only some of the linked entities are on. It reads the saved Active state, so flipping the switch without saving does not move it. See Serving state.

Share — giving another user access to your entity. Three types: Is Owner, Can Edit and Can View. See Shares.

Show tracking URLs — the per-creative list of impression and click URLs for a placement. A tab on the placement’s Ad tags & API card in the new interface, a header button in the classic one. It is the only serving surface a Click-through placement has. See Invocation code types.

Site — the top supply-side entity: a publisher website or app property. Hierarchy: Site → Zone → Placement.

Split columns by — the multi-select that spreads dimensions across a report’s columns rather than down its rows. Most entries combine; Custom parameter is the one exclusive option. See Analytics metrics.

SSP — supply-side platform. Epom’s role when it asks DSPs to bid.

Statistic type — which report you are running. Seven values exist, six can be selected, and the General report is none of them — it is what you get by omitting the parameter. See Analytics reports via API.

Sub-ID — a correlation identifier passed on the click and returned in the S2S postback to attribute a conversion without cookies.

Supply — the publisher side of the network: sites, zones, placements.

T

Targeting — demand-side rules deciding which traffic an advertiser, campaign or banner serves to. The platform defines 43 rule types: 38 are dimensions you pick from the add-rule list, 40 have a REST create endpoint, and 5 are blocked-entity types no picker offers. The supply-side equivalent is pretargeting. See Targeting dimensions.

Template — two unrelated features share the name: a reusable rich-media creative skeleton with configurable properties, which is the basis of Template banners; and the admin email/SMS notification templates.

Traffic quality — the fraud-scoring area of the account: per-inventory risk scores, four levels and downloadable scan reports. The levels are advisory — see Blocked. See Moderation & traffic quality.

Traffic source — an analytics and targeting dimension naming the integration a request arrived through, not anything about the visitor. It has exactly one value today, HEADER_BIDDING; an empty cell means the ordinary ad tag.

Transaction ID — the per-impression identifier that ties an ad request to everything downstream: it is what imp.id carries in a bid request, what the raw logs record, and what $$TRANSACTION_ID_STR$$ and $$TRANSACTION_ID_LONG$$ resolve to. Not the placement id.

Trees and grids — the two ways an entity workspace renders. Trees give an expandable hierarchy with drag-and-drop re-parenting and the action rail; grids show one level at a time with server-driven sort and paging. Your default is a saved preference; the workspace toggle overrides it for the session only. See UI navigation.

U

Unattributed — a conversion that arrived with no attribution type declared and none could be inferred. Also the default value of an action’s Default attribution type.

Universal JS — the asynchronous invocation code: an <ins> element plus a loader <script>. Its code and API name is AJAX, so searching for the label finds nothing. See Invocation code types.

V

VAST — video ad serving template, the XML format a player fetches to get a video ad. Epom generates VAST tag URLs at versions 2.0 through 4.3, defaulting to 4.2. See Video, VAST & VPAID.

VMAP — video multiple ad playlist, the schedule that tells a player when to run ad breaks. Its tag URL is pinned server-side and does not follow the VAST version picker.

VPAID — video player ad-serving interface definition, the interactive-video runtime. Its payload is code that runs in the player, which makes it less transparent than plain VAST. See VAST vs VPAID.

W

Waterfall — priority-ordered banner fallback: all banners of a higher tier are tried before lower tiers, ending in the passback. Also called daisy chaining; the API calls the configuration chaining.

White label — running the platform under your own brand: domain, logo, theme colors and feature toggles. See White-label.

Z

Zone — the middle supply-side level: a named group of placements inside a site (for example “Homepage” or “Mobile web”).

Related

• Analytics metrics — the money columns and every ratio, in full
• Analytics dimensions — what you can group and filter by
• Banner types and Placement types
• Capping matrix and Targeting dimensions
• Statuses & indicators — every glyph, pill and badge
• UI navigation, grids & trees
• REST API overview