Skip to content

:hedgehog: PostHog is the leading platform for building self-driving products. Our developer tools – AI observability, analytics, session replay, flags, experiments, error tracking, logs, and more – capture all the context agents need to diagnose problems, uncover opportunities, and ship fixes. Steer it all from Slack, web, desktop, or the MCP.

feat(marketing): add interactive ad performance sparklines (#120441)

Javier Bahamondes653769f607e4
diff --git a/docs/internal/marketing-source-suggestions.md b/docs/internal/marketing-source-suggestions.md
index 313002bba4d9..4e7055a25e2f 100644
--- a/docs/internal/marketing-source-suggestions.md
+++ b/docs/internal/marketing-source-suggestions.md
@@ -301,3 +301,13 @@ With source onboarding disabled, the existing banner remains available.
The Ad performance summary uses the same metric cards as Web analytics.
Cards preserve period comparisons, currency formatting, and missing-value labels, and wrap to the available scene width.
+
+Ad performance cards include time-series sparklines for advertising metrics, including platform-reported conversions.
+Select an advertising metric card to update the existing trend chart while keeping the date range, interval, and source filters.
+CPC, CTR, reported ROAS, and cost per reported conversion use ratios of bucket totals in both sparklines and the main trend, with monetary values converted to the project currency; undefined ratios leave gaps rather than zeroes.
+Attributed PostHog conversion goals retain their summary cards without sparklines or selection.
+
+Advertising metrics appear first, followed by attributed conversion goals and their cost per conversion.
+The goal cards use a separate grid so cards without sparklines keep a compact height.
+
+The overview loads on demand so other query surfaces do not load marketing-specific validation eagerly.
diff --git a/frontend/snapshots.yml b/frontend/snapshots.yml
index 538e58dd863d..6dc9393f78f2 100644
--- a/frontend/snapshots.yml
+++ b/frontend/snapshots.yml
@@ -6036,6 +6036,10 @@ snapshots:
hash: v1.k794b7964.f9059da612f8cb309c1100982a7352bf0c010c585701ed9f703ed1962da1a8b2.4Wbe0lwcRKZju4mbcJDVTwzvT2DtSSqF1VwdFStAA5o
logs-logsretentionperiodpicker--stored-custom-months-without-paid-retention--light:
hash: v1.k794b7964.d8d79824779beb9150246c932e4f2b756133d02aca4ca269271b98d0e9082480.PCD7I0jw8Fis5NRmUgRrjvUpwz5Dap-BO1UKMSJu5Ns
+ marketing-analytics-ad-performance-metric-trends--with-trends--dark:
+ hash: v1.k794b7964.9a2823d7597aac44d396018c51abba4545f3079ea859dc75855cb064add75317.2DtYQx5epqWzQwQS5oFiY-QYBMVzZoDVZZbr7SPYrsk
+ marketing-analytics-ad-performance-metric-trends--with-trends--light:
+ hash: v1.k794b7964.2f4ccaa52f1dac23d7792a6c737b167c3327c76418e35ae15289353f90d3010d.fL2tqTuwK5Zjhq1FOUJkaXAnBHs6qKnexB-SWPfu7rM
marketing-analytics-add-source--dismissed-everywhere--dark:
hash: v1.k794b7964.f23fe271dbeafe8d745237090a5a115e167653cdb000e74a3245efd986450ff0.ymRCgme2xTcwSDlK8Mjxj_zM6bDHbYjLGl0ncGYQfuo
marketing-analytics-add-source--dismissed-everywhere--light:
diff --git a/frontend/src/queries/Query/Query.tsx b/frontend/src/queries/Query/Query.tsx
index 73cc8cf9c1c5..1204363b398f 100644
--- a/frontend/src/queries/Query/Query.tsx
+++ b/frontend/src/queries/Query/Query.tsx
@@ -5,7 +5,6 @@ import { LemonDivider } from 'lib/lemon-ui/LemonDivider'
import { Spinner } from 'lib/lemon-ui/Spinner'
import { lazyWithRetry } from 'lib/utils/retryImport'
import { HogDebug } from 'scenes/debug/HogDebug'
-import { MarketingAnalyticsOverview } from 'scenes/web-analytics/tabs/marketing-analytics/frontend/components/MarketingAnalyticsOverview/MarketingAnalyticsOverview'
import { ErrorBoundary } from '~/layout/ErrorBoundary'
import { DataNode } from '~/queries/nodes/DataNode/DataNode'
@@ -46,6 +45,11 @@ import {
isWebVitalsQuery,
} from '../utils'
+const MarketingAnalyticsOverview = lazyWithRetry(() =>
+ import('scenes/web-analytics/tabs/marketing-analytics/frontend/components/MarketingAnalyticsOverview/MarketingAnalyticsOverview').then(
+ (m) => ({ default: m.MarketingAnalyticsOverview })
+ )
+)
const WebVitals = lazyWithRetry(() =>
import('~/queries/nodes/WebVitals/WebVitals').then((m) => ({ default: m.WebVitals }))
)
@@ -236,13 +240,15 @@ export function Query<Q extends Node>(props: QueryProps<Q>): JSX.Element | null
)
} else if (isMarketingAnalyticsAggregatedQuery(query)) {
component = (
- <MarketingAnalyticsOverview
- attachTo={props.attachTo}
- query={query}
- cachedResults={props.cachedResults}
- context={queryContext}
- uniqueKey={uniqueKey}
- />
+ <Suspense fallback={<Spinner />}>
+ <MarketingAnalyticsOverview
+ attachTo={props.attachTo}
+ query={query}
+ cachedResults={props.cachedResults}
+ context={queryContext}
+ uniqueKey={uniqueKey}
+ />
+ </Suspense>
)
} else if (isWebVitalsQuery(query)) {
component = (
diff --git a/frontend/src/queries/nodes/OverviewGrid/OverviewMetricCardGrid.tsx b/frontend/src/queries/nodes/OverviewGrid/OverviewMetricCardGrid.tsx
index 44babf760123..d713ba9284b4 100644
--- a/frontend/src/queries/nodes/OverviewGrid/OverviewMetricCardGrid.tsx
+++ b/frontend/src/queries/nodes/OverviewGrid/OverviewMetricCardGrid.tsx
@@ -3,10 +3,11 @@ import type React from 'react'
import { IconWarning } from '@posthog/icons'
import { Link } from '@posthog/lemon-ui'
-import { type MetricChange } from '@posthog/quill-charts'
+import { type ChartTheme, type MetricChange } from '@posthog/quill-charts'
import { AnalyticsMetricCard } from 'lib/components/AnalyticsMetricCard/AnalyticsMetricCard'
import { PreAggregatedBadge } from 'lib/components/PreAggregatedBadge'
+import { dayjs } from 'lib/dayjs'
import { Tooltip } from 'lib/lemon-ui/Tooltip'
import { range } from 'lib/utils/arrays'
import { cn } from 'lib/utils/css-classes'
@@ -16,7 +17,7 @@ import { WebAnalyticsPreComputeStrategy } from '~/queries/schema/schema-general'
import { formatItem, NO_BASELINE_CHANGE_SENTINEL, OverviewItem, SamplingNotice, SamplingRate } from './OverviewGrid'
-export type OverviewMetricCardItem = OverviewItem
+export type OverviewMetricCardItem = OverviewItem & { sparkline?: { data: number[]; labels: string[] } }
interface OverviewMetricCardGridProps {
items: OverviewMetricCardItem[]
@@ -25,6 +26,7 @@ interface OverviewMetricCardGridProps {
numSkeletons: number
emptyMessage?: React.ReactNode
className?: string
+ sparklineTheme?: ChartTheme
samplingRate?: SamplingRate
preComputeStrategy?: WebAnalyticsPreComputeStrategy
onDisablePrecompute?: () => void
@@ -38,6 +40,7 @@ export function OverviewMetricCardGrid({
numSkeletons,
emptyMessage,
className,
+ sparklineTheme,
samplingRate,
preComputeStrategy,
onDisablePrecompute,
@@ -57,6 +60,7 @@ export function OverviewMetricCardGrid({
<MetricCardCell
key={item.key}
item={item}
+ sparklineTheme={sparklineTheme}
emptyMessage={emptyMessage}
stretch={layout === 'contents'}
preComputeStrategy={preComputeStrategy}
@@ -72,6 +76,7 @@ export function OverviewMetricCardGrid({
function MetricCardCell({
item,
+ sparklineTheme,
emptyMessage,
stretch,
preComputeStrategy,
@@ -79,6 +84,7 @@ function MetricCardCell({
labelFromKey,
}: {
item: OverviewMetricCardItem
+ sparklineTheme?: ChartTheme
emptyMessage?: React.ReactNode
stretch?: boolean
preComputeStrategy?: WebAnalyticsPreComputeStrategy
@@ -105,7 +111,14 @@ function MetricCardCell({
formatValue={(value) =>
formatItem(typeof item.value === 'number' ? value : item.value, item.kind, { currency: baseCurrency })
}
- subtitle={subtitle}
+ subtitle={item.sparkline && sparklineTheme ? undefined : subtitle}
+ restingSubtitle={subtitle}
+ data={sparklineTheme ? item.sparkline?.data : undefined}
+ labels={item.sparkline?.labels}
+ theme={sparklineTheme}
+ formatLabel={(label) => dayjs(label).format('MMM D')}
+ sparklineHeight={48}
+ sparklineClassName="mt-3"
adornment={
preComputeStrategy === WebAnalyticsPreComputeStrategy.LazyPrecompute ? (
<PreAggregatedBadge variant="precomputed" position="bottom-right" onDisable={onDisablePrecompute} />
diff --git a/frontend/src/scenes/web-analytics/tabs/marketing-analytics/frontend/components/MarketingAnalyticsOverview/MarketingAnalyticsOverview.test.tsx b/frontend/src/scenes/web-analytics/tabs/marketing-analytics/frontend/components/MarketingAnalyticsOverview/MarketingAnalyticsOverview.test.tsx
new file mode 100644
index 000000000000..02965d88f016
--- /dev/null
+++ b/frontend/src/scenes/web-analytics/tabs/marketing-analytics/frontend/components/MarketingAnalyticsOverview/MarketingAnalyticsOverview.test.tsx
@@ -0,0 +1,48 @@
+import { cleanup, fireEvent, render, screen } from '@testing-library/react'
+
+import { marketingAnalyticsLogic } from 'scenes/web-analytics/tabs/marketing-analytics/frontend/logic/marketingAnalyticsLogic'
+import { marketingAnalyticsTilesLogic } from 'scenes/web-analytics/tabs/marketing-analytics/frontend/logic/marketingAnalyticsTilesLogic'
+
+import { Query } from '~/queries/Query/Query'
+import { InsightVizNode, MarketingAnalyticsAggregatedQueryResponse, NodeKind } from '~/queries/schema/schema-general'
+import { initKeaTests } from '~/test/init'
+
+const response: MarketingAnalyticsAggregatedQueryResponse = {
+ results: {
+ Cost: { key: 'Cost', value: 100, kind: 'currency' },
+ Clicks: { key: 'Clicks', value: 50, kind: 'unit' },
+ CPC: { key: 'CPC', value: 2, kind: 'currency' },
+ CTR: { key: 'CTR', value: 5, kind: 'percentage' },
+ 'Trial started': { key: 'Trial started', value: 10, kind: 'unit' },
+ },
+}
+
+describe('Ad performance metric selection', () => {
+ beforeEach(() => initKeaTests())
+ afterEach(cleanup)
+
+ it.each([
+ ['Clicks', 'clicks'],
+ ['CPC', 'cpc'],
+ ['CTR', 'ctr'],
+ ])('selecting %s updates the existing trend without changing the filters', async (label, metric) => {
+ render(
+ <Query
+ query={{ kind: NodeKind.MarketingAnalyticsAggregatedQuery, properties: [] }}
+ cachedResults={response}
+ context={{}}
+ uniqueKey="metric-selection"
+ />
+ )
+ const button = (await screen.findByText(label)).closest('button')!
+ const before = (marketingAnalyticsTilesLogic.values.marketingChartTile.query as InsightVizNode).source
+ fireEvent.click(button)
+ expect(marketingAnalyticsLogic.values.tileColumnSelection).toBe(metric)
+ expect(button.getAttribute('aria-pressed')).toBe('true')
+ expect(marketingAnalyticsTilesLogic.values.marketingChartTile.title).toBe(`Marketing ${metric}`)
+ expect(
+ (marketingAnalyticsTilesLogic.values.marketingChartTile.query as InsightVizNode).source.dateRange
+ ).toEqual(before.dateRange)
+ expect(screen.getByText('Trial Started').closest('button')).toBeNull()
+ })
+})
diff --git a/frontend/src/scenes/web-analytics/tabs/marketing-analytics/frontend/components/MarketingAnalyticsOverview/MarketingAnalyticsOverview.tsx b/frontend/src/scenes/web-analytics/tabs/marketing-analytics/frontend/components/MarketingAnalyticsOverview/MarketingAnalyticsOverview.tsx
index 584cf61d7ada..1865ec87c4e0 100644
--- a/frontend/src/scenes/web-analytics/tabs/marketing-analytics/frontend/components/MarketingAnalyticsOverview/MarketingAnalyticsOverview.tsx
+++ b/frontend/src/scenes/web-analytics/tabs/marketing-analytics/frontend/components/MarketingAnalyticsOverview/MarketingAnalyticsOverview.tsx
@@ -1,6 +1,7 @@
-import { BuiltLogic, LogicWrapper, useValues } from 'kea'
+import { BuiltLogic, LogicWrapper, useActions, useValues } from 'kea'
import { useMemo, useState } from 'react'
+import { useChartTheme } from 'lib/charts/hooks'
import { useAttachedLogic } from 'lib/logic/scenes/useAttachedLogic'
import { InsightErrorState } from 'scenes/insights/EmptyStates'
@@ -14,7 +15,15 @@ import {
import { QueryContext } from '~/queries/types'
import { MarketingAnalyticsNotReady } from '~/scenes/marketing-analytics/MarketingAnalyticsNotReady'
+import { adPerformanceMetricKey } from 'products/marketing_analytics/frontend/adPerformanceMetrics'
+import { adPerformanceTrendsLogic } from 'products/marketing_analytics/frontend/adPerformanceTrendsLogic'
+
+import { marketingAnalyticsLogic } from '../../logic/marketingAnalyticsLogic'
import { marketingAnalyticsSettingsLogic } from '../../logic/marketingAnalyticsSettingsLogic'
+import {
+ MARKETING_ANALYTICS_DATA_COLLECTION_NODE_ID,
+ marketingAnalyticsTilesLogic,
+} from '../../logic/marketingAnalyticsTilesLogic'
import {
MarketingAnalyticsValidationWarningBanner,
validateConversionGoals,
@@ -48,19 +57,36 @@ export function MarketingAnalyticsOverview(props: {
const marketingOverviewQueryResponse = response as MarketingAnalyticsAggregatedQueryResponse | undefined
+ const { createAdPerformanceSparklineQuery } = useValues(marketingAnalyticsTilesLogic)
+ const { tileColumnSelection } = useValues(marketingAnalyticsLogic)
+ const { setTileColumnSelection } = useActions(marketingAnalyticsLogic)
+ const { sparklines } = useValues(
+ adPerformanceTrendsLogic({
+ query: createAdPerformanceSparklineQuery(props.query),
+ queryKey: `${key}-sparklines`,
+ dataNodeCollectionId: dataNodeCollectionId ?? MARKETING_ANALYTICS_DATA_COLLECTION_NODE_ID,
+ })
+ )
+ const sparklineTheme = useChartTheme()
const samplingRate = marketingOverviewQueryResponse?.samplingRate
const validationWarnings = useMemo(() => validateConversionGoals(conversion_goals), [conversion_goals])
const overviewItems: OverviewMetricCardItem[] = marketingOverviewQueryResponse?.results
- ? Object.entries(marketingOverviewQueryResponse.results).map(([key, item]) => ({
- key,
- value: item.value,
- previous: item.previous,
- changeFromPreviousPct: item.changeFromPreviousPct,
- kind: item.kind,
- isIncreaseBad: item.isIncreaseBad,
- }))
+ ? Object.entries(marketingOverviewQueryResponse.results).map(([key, item]) => {
+ const metric = adPerformanceMetricKey(key)
+ return {
+ key,
+ value: item.value,
+ previous: item.previous,
+ changeFromPreviousPct: item.changeFromPreviousPct,
+ kind: item.kind,
+ isIncreaseBad: item.isIncreaseBad,
+ selected: metric ? tileColumnSelection === metric : undefined,
+ onClick: metric ? () => setTileColumnSelection(metric) : undefined,
+ sparkline: metric && typeof item.value === 'number' ? sparklines[metric] : undefined,
+ }
+ })
: []
// Calculate number of skeletons based on expected metrics
@@ -87,15 +113,28 @@ export function MarketingAnalyticsOverview(props: {
return (
<>
{allWarnings.length > 0 && <MarketingAnalyticsValidationWarningBanner warnings={allWarnings} />}
- <OverviewMetricCardGrid
- className="grid-cols-[repeat(auto-fit,minmax(12rem,1fr))]"
- items={overviewItems.filter(filterEmptyMetrics)}
- loading={responseLoading}
- numSkeletons={numSkeletons}
- emptyMessage="-"
- samplingRate={samplingRate}
- labelFromKey={(key) => <span className="[overflow-wrap:anywhere]">{labelFromKey(key)}</span>}
- />
+ <div className="flex flex-col gap-2">
+ {[
+ overviewItems.filter((item) => adPerformanceMetricKey(item.key)),
+ overviewItems.filter((item) => !adPerformanceMetricKey(item.key)),
+ ].map((items, index) =>
+ (responseLoading && index === 0) || (!responseLoading && items.length > 0) ? (
+ <OverviewMetricCardGrid
+ key={index}
+ sparklineTheme={sparklineTheme}
+ className="grid-cols-[repeat(auto-fit,minmax(12rem,1fr))]"
+ items={items.filter(filterEmptyMetrics)}
+ loading={responseLoading}
+ numSkeletons={numSkeletons}
+ emptyMessage="-"
+ samplingRate={index === 0 ? samplingRate : undefined}
+ labelFromKey={(key) => (
+ <span className="[overflow-wrap:anywhere]">{labelFromKey(key)}</span>
+ )}
+ />
+ ) : null
+ )}
+ </div>
</>
)
}
diff --git a/frontend/src/scenes/web-analytics/tabs/marketing-analytics/frontend/logic/__snapshots__/utils.test.ts.snap b/frontend/src/scenes/web-analytics/tabs/marketing-analytics/frontend/logic/__snapshots__/utils.test.ts.snap
index 9fca5db009e9..bf7887d26f3b 100644
--- a/frontend/src/scenes/web-analytics/tabs/marketing-analytics/frontend/logic/__snapshots__/utils.test.ts.snap
+++ b/frontend/src/scenes/web-analytics/tabs/marketing-analytics/frontend/logic/__snapshots__/utils.test.ts.snap
@@ -398,7 +398,7 @@ exports[`marketing analytics utils createMarketingTile snapshots BingAds - roas
"id_field": "campaign_id",
"kind": "DataWarehouseNode",
"math": "hogql",
- "math_hogql": "SUM(ifNull(toFloat(revenue), 0)) / nullIf(SUM(ifNull(toFloat(spend), 0)), 0)",
+ "math_hogql": "SUM(toFloat(convertCurrency(coalesce(currency_code, 'USD'), 'USD', ifNull(toFloat(revenue), 0)))) / nullIf(SUM(toFloat(convertCurrency(coalesce(currency_code, 'USD'), 'USD', ifNull(toFloat(spend), 0)))), 0)",
"name": "bing",
"table_name": "prefix.campaign_performance_report",
"timestamp_field": "time_period",
@@ -533,7 +533,7 @@ exports[`marketing analytics utils createMarketingTile snapshots GoogleAds - roa
"id_field": "campaign_id",
"kind": "DataWarehouseNode",
"math": "hogql",
- "math_hogql": "SUM(ifNull(toFloat(metrics_conversions_value), 0)) / nullIf(SUM(ifNull(toFloat(metrics_cost_micros), 0) / 1000000), 0)",
+ "math_hogql": "SUM(toFloat(convertCurrency(coalesce(customer_currency_code, 'USD'), 'USD', ifNull(toFloat(metrics_conversions_value), 0)))) / nullIf(SUM(toFloat(convertCurrency(coalesce(customer_currency_code, 'USD'), 'USD', ifNull(toFloat(metrics_cost_micros), 0) / 1000000))), 0)",
"name": "google",
"table_name": "prefix.campaign_overview_stats",
"timestamp_field": "segments_date",
@@ -668,7 +668,7 @@ exports[`marketing analytics utils createMarketingTile snapshots LinkedinAds - r
"id_field": "campaign_group_id",
"kind": "DataWarehouseNode",
"math": "hogql",
- "math_hogql": "SUM(ifNull(toFloat(conversion_value_in_local_currency), 0)) / nullIf(SUM(ifNull(toFloat(cost_in_usd), 0)), 0)",
+ "math_hogql": "toFloat(convertCurrency('USD', 'USD', SUM(ifNull(toFloat(conversion_value_in_local_currency), 0)))) / nullIf(toFloat(convertCurrency('USD', 'USD', SUM(ifNull(toFloat(cost_in_usd), 0)))), 0)",
"name": "linkedin",
"table_name": "prefix.campaign_group_stats",
"timestamp_field": "date_start",
@@ -803,7 +803,7 @@ exports[`marketing analytics utils createMarketingTile snapshots MetaAds - roas
"id_field": "campaign_id",
"kind": "DataWarehouseNode",
"math": "hogql",
- "math_hogql": "SUM(if(arraySum(x -> JSONExtractFloat(x, 'value'), arrayFilter(x -> JSONExtractString(x, 'action_type') IN ('omni_purchase', 'omni_lead', 'omni_complete_registration', 'omni_app_install', 'omni_subscribe'), JSONExtractArrayRaw(coalesce(action_values, '[]')))) > 0, arraySum(x -> JSONExtractFloat(x, 'value'), arrayFilter(x -> JSONExtractString(x, 'action_type') IN ('omni_purchase', 'omni_lead', 'omni_complete_registration', 'omni_app_install', 'omni_subscribe'), JSONExtractArrayRaw(coalesce(action_values, '[]')))), arraySum(x -> JSONExtractFloat(x, 'value'), arrayFilter(x -> JSONExtractString(x, 'action_type') IN ('purchase', 'lead', 'complete_registration', 'app_install', 'subscribe'), JSONExtractArrayRaw(coalesce(action_values, '[]')))))) / nullIf(SUM(ifNull(toFloat(spend), 0)), 0)",
+ "math_hogql": "SUM(toFloat(convertCurrency(coalesce(account_currency, 'USD'), 'USD', if(arraySum(x -> JSONExtractFloat(x, 'value'), arrayFilter(x -> JSONExtractString(x, 'action_type') IN ('omni_purchase', 'omni_lead', 'omni_complete_registration', 'omni_app_install', 'omni_subscribe'), JSONExtractArrayRaw(coalesce(action_values, '[]')))) > 0, arraySum(x -> JSONExtractFloat(x, 'value'), arrayFilter(x -> JSONExtractString(x, 'action_type') IN ('omni_purchase', 'omni_lead', 'omni_complete_registration', 'omni_app_install', 'omni_subscribe'), JSONExtractArrayRaw(coalesce(action_values, '[]')))), arraySum(x -> JSONExtractFloat(x, 'value'), arrayFilter(x -> JSONExtractString(x, 'action_type') IN ('purchase', 'lead', 'complete_registration', 'app_install', 'subscribe'), JSONExtractArrayRaw(coalesce(action_values, '[]')))))))) / nullIf(SUM(toFloat(convertCurrency(coalesce(account_currency, 'USD'), 'USD', ifNull(toFloat(spend), 0)))), 0)",
"name": "meta",
"table_name": "prefix.campaign_stats",
"timestamp_field": "date_stop",
@@ -1073,7 +1073,7 @@ exports[`marketing analytics utils createMarketingTile snapshots PinterestAds -
"id_field": "campaign_id",
"kind": "DataWarehouseNode",
"math": "hogql",
- "math_hogql": "SUM(ifNull(toFloat(total_checkout_value_in_micro_dollar), 0) / 1000000) / nullIf(SUM(ifNull(toFloat(spend_in_dollar), 0)), 0)",
+ "math_hogql": "SUM(toFloat(convertCurrency(coalesce(currency, 'USD'), 'USD', ifNull(toFloat(total_checkout_value_in_micro_dollar), 0) / 1000000))) / nullIf(SUM(toFloat(convertCurrency(coalesce(currency, 'USD'), 'USD', ifNull(toFloat(spend_in_dollar), 0)))), 0)",
"name": "pinterest",
"table_name": "prefix.campaign_analytics",
"timestamp_field": "date",
@@ -1208,7 +1208,7 @@ exports[`marketing analytics utils createMarketingTile snapshots RedditAds - roa
"id_field": "campaign_id",
"kind": "DataWarehouseNode",
"math": "hogql",
- "math_hogql": "SUM(ifNull(toFloat(conversion_purchase_total_value), 0) / 100 + ifNull(toFloat(conversion_signup_total_value), 0) / 100) / nullIf(SUM(ifNull(toFloat(spend), 0) / 1000000), 0)",
+ "math_hogql": "SUM(toFloat(convertCurrency(coalesce(currency, 'USD'), 'USD', ifNull(toFloat(conversion_purchase_total_value), 0) / 100 + ifNull(toFloat(conversion_signup_total_value), 0) / 100))) / nullIf(SUM(toFloat(convertCurrency(coalesce(currency, 'USD'), 'USD', ifNull(toFloat(spend), 0) / 1000000))), 0)",
"name": "reddit",
"table_name": "prefix.campaign_report",
"timestamp_field": "date",
@@ -1478,7 +1478,7 @@ exports[`marketing analytics utils createMarketingTile snapshots SnapchatAds - r
"id_field": "id",
"kind": "DataWarehouseNode",
"math": "hogql",
- "math_hogql": "SUM(ifNull(toFloat(conversion_purchases_value), 0) + ifNull(toFloat(conversion_sign_ups_value), 0) + ifNull(toFloat(conversion_subscribe_value), 0)) / nullIf(SUM(ifNull(toFloat(spend), 0) / 1000000), 0)",
+ "math_hogql": "SUM(toFloat(convertCurrency(coalesce(currency, 'USD'), 'USD', ifNull(toFloat(conversion_purchases_value), 0) + ifNull(toFloat(conversion_sign_ups_value), 0) + ifNull(toFloat(conversion_subscribe_value), 0)))) / nullIf(SUM(toFloat(convertCurrency(coalesce(currency, 'USD'), 'USD', ifNull(toFloat(spend), 0) / 1000000))), 0)",
"name": "snapchat",
"table_name": "prefix.campaign_stats_daily",
"timestamp_field": "start_time",
@@ -1613,7 +1613,7 @@ exports[`marketing analytics utils createMarketingTile snapshots TikTokAds - roa
"id_field": "campaign_id",
"kind": "DataWarehouseNode",
"math": "hogql",
- "math_hogql": "SUM(ifNull(toFloat(total_complete_payment_value), 0)) / nullIf(SUM(ifNull(toFloat(spend), 0)), 0)",
+ "math_hogql": "SUM(toFloat(convertCurrency(coalesce(currency, 'USD'), 'USD', ifNull(toFloat(total_complete_payment_value), 0)))) / nullIf(SUM(toFloat(convertCurrency(coalesce(currency, 'USD'), 'USD', ifNull(toFloat(spend), 0)))), 0)",
"name": "tiktok",
"table_name": "prefix.campaign_report",
"timestamp_field": "stat_time_day",
diff --git a/frontend/src/scenes/web-analytics/tabs/marketing-analytics/frontend/logic/marketingAnalyticsTilesLogic.ts b/frontend/src/scenes/web-analytics/tabs/marketing-analytics/frontend/logic/marketingAnalyticsTilesLogic.ts
index 4ac16c1e72c7..d65ab6cd7e4a 100644
--- a/frontend/src/scenes/web-analytics/tabs/marketing-analytics/frontend/logic/marketingAnalyticsTilesLogic.ts
+++ b/frontend/src/scenes/web-analytics/tabs/marketing-analytics/frontend/logic/marketingAnalyticsTilesLogic.ts
@@ -18,6 +18,8 @@ import {
MARKETING_ANALYTICS_DRILL_DOWN_CONFIG,
MARKETING_ANALYTICS_SCHEMA,
MarketingAnalyticsBaseColumns,
+ MarketingAnalyticsAggregatedQuery,
+ TrendsQuery,
MarketingAnalyticsConstants,
MarketingAnalyticsDrillDownLevel,
MarketingAnalyticsTableQuery,
@@ -35,7 +37,9 @@ import {
PropertyOperator,
} from '~/types'
-import { marketingAnalyticsLogic } from './marketingAnalyticsLogic'
+import { buildAdPerformanceTrendsQuery } from 'products/marketing_analytics/frontend/adPerformanceMetrics'
+
+import { marketingAnalyticsLogic, type NativeSource, type ExternalTable } from './marketingAnalyticsLogic'
import { marketingAnalyticsTableLogic } from './marketingAnalyticsTableLogic'
import {
getOrderBy,
@@ -49,11 +53,13 @@ import {
export const MARKETING_ANALYTICS_DATA_COLLECTION_NODE_ID = pageCollectionId('marketing-analytics')
const isSchemaBackedMarketingColumn = (column: validColumnsForTiles): column is rawColumnsForTiles =>
- column !== 'roas' && column !== 'cost_per_reported_conversion'
+ column !== 'roas' && column !== 'cost_per_reported_conversion' && column !== 'cpc' && column !== 'ctr'
// Maps a chart column to a HogQL aggregate over the native cost precompute table. Costs are already
// converted to the team's base currency at materialization time, so no convertCurrency() is needed.
const PRECOMPUTE_METRIC_EXPR: Record<validColumnsForTiles, string> = {
+ cpc: 'sum(cost) / nullIf(sum(clicks), 0)',
+ ctr: 'sum(clicks) / nullIf(sum(impressions), 0) * 100',
cost: 'sum(cost)',
clicks: 'sum(clicks)',
impressions: 'sum(impressions)',
@@ -90,10 +96,13 @@ export interface marketingAnalyticsTilesLogicValues {
shouldFilterTestAccounts: boolean // marketingAnalyticsLogic
sourcesInitialLoading: boolean // marketingAnalyticsLogic
tileColumnSelection: validColumnsForTiles // marketingAnalyticsLogic
+ validExternalTables: ExternalTable[] // marketingAnalyticsLogic
+ validNativeSources: NativeSource[] // marketingAnalyticsLogic
defaultColumns: string[] // marketingAnalyticsTableLogic
query: DataTableNode | null // marketingAnalyticsTableLogic
campaignBreakdownTile: QueryTile | null
campaignCostsBreakdown: DataTableNode | null
+ createAdPerformanceSparklineQuery: (query: MarketingAnalyticsAggregatedQuery) => TrendsQuery
marketingChartTile: QueryTile
overviewTile: QueryTile
tiles: QueryTile[]
@@ -102,6 +111,17 @@ export interface marketingAnalyticsTilesLogicValues {
// Generated by kea-typegen. Update if you're an agent, ignore if you're human.
export interface marketingAnalyticsTilesLogicMeta {
__keaTypeGenInternalSelectorTypes: {
+ createAdPerformanceSparklineQuery: (
+ dateFilter: {
+ dateFrom: string | null
+ dateTo: string | null
+ interval: IntervalType
+ },
+ validNativeSources: NativeSource[],
+ validExternalTables: ExternalTable[],
+ baseCurrency: CurrencyCode,
+ featureFlags: FeatureFlagsSet
+ ) => (query: MarketingAnalyticsAggregatedQuery) => TrendsQuery
overviewTile: (
compareFilter: CompareFilter,
dateFilter: {
@@ -184,6 +204,8 @@ export const marketingAnalyticsTilesLogic = kea<marketingAnalyticsTilesLogicType
'drillDownLevel',
'baseCurrency',
'includeConversionGoals',
+ 'validNativeSources',
+ 'validExternalTables',
],
marketingAnalyticsTableLogic,
['query', 'defaultColumns'],
@@ -192,6 +214,25 @@ export const marketingAnalyticsTilesLogic = kea<marketingAnalyticsTilesLogicType
],
})),
selectors({
+ createAdPerformanceSparklineQuery: [
+ (s) => [s.dateFilter, s.validNativeSources, s.validExternalTables, s.baseCurrency, s.featureFlags],
+ (
+ dateFilter: { interval: IntervalType },
+ nativeSources: NativeSource[],
+ externalTables: ExternalTable[],
+ baseCurrency: CurrencyCode,
+ featureFlags: FeatureFlagsSet
+ ): ((query: MarketingAnalyticsAggregatedQuery) => TrendsQuery) =>
+ (query) =>
+ buildAdPerformanceTrendsQuery(
+ query,
+ dateFilter.interval,
+ nativeSources,
+ externalTables,
+ baseCurrency,
+ !!featureFlags[FEATURE_FLAGS.MARKETING_ANALYTICS_COSTS_PRECOMPUTATION]
+ ),
+ ],
// One selector per tile so an input change only invalidates the tile that uses
// it — entries keep identity in `tiles` below, so `dataNodeLogic` doesn't refetch.
overviewTile: [
@@ -262,9 +303,10 @@ export const marketingAnalyticsTilesLogic = kea<marketingAnalyticsTilesLogicType
const hasSources = createMarketingDataWarehouseNodes.length > 0
const metricExpr = PRECOMPUTE_METRIC_EXPR[tileColumnSelection] ?? 'sum(cost)'
const isCurrency =
- tileColumnSelection && isSchemaBackedMarketingColumn(tileColumnSelection)
+ ['cpc', 'cost_per_reported_conversion'].includes(tileColumnSelection) ||
+ (tileColumnSelection && isSchemaBackedMarketingColumn(tileColumnSelection)
? MARKETING_ANALYTICS_SCHEMA[tileColumnSelection].isCurrency
- : false
+ : false)
const { symbol: currencySymbol, isPrefix: currencyIsPrefix } = getCurrencySymbol(baseCurrency)
// Gated behind the same flag as the backend cost precompute read-side: when off, the tile
@@ -348,7 +390,12 @@ export const marketingAnalyticsTilesLogic = kea<marketingAnalyticsTilesLogicType
display: chartDisplayType,
aggregationAxisFormat: 'numeric',
aggregationAxisPrefix: isCurrency && currencyIsPrefix ? currencySymbol : undefined,
- aggregationAxisPostfix: isCurrency && !currencyIsPrefix ? ` ${currencySymbol}` : undefined,
+ aggregationAxisPostfix:
+ tileColumnSelection === 'ctr'
+ ? '%'
+ : isCurrency && !currencyIsPrefix
+ ? ` ${currencySymbol}`
+ : undefined,
},
tags: MARKETING_ANALYTICS_DEFAULT_QUERY_TAGS,
},
diff --git a/frontend/src/scenes/web-analytics/tabs/marketing-analytics/frontend/logic/marketingCostTile.ts b/frontend/src/scenes/web-analytics/tabs/marketing-analytics/frontend/logic/marketingCostTile.ts
index a77d0ff522da..9cc377e3fb0e 100644
--- a/frontend/src/scenes/web-analytics/tabs/marketing-analytics/frontend/logic/marketingCostTile.ts
+++ b/frontend/src/scenes/web-analytics/tabs/marketing-analytics/frontend/logic/marketingCostTile.ts
@@ -1,4 +1,9 @@
-import { DataWarehouseNode, MarketingAnalyticsConstants, NodeKind } from '~/queries/schema/schema-general'
+import {
+ DataWarehouseNode,
+ MarketingAnalyticsColumnsSchemaNames,
+ MarketingAnalyticsConstants,
+ NodeKind,
+} from '~/queries/schema/schema-general'
import { HogQLMathType } from '~/types'
import { ExternalTable } from './marketingAnalyticsLogic'
@@ -59,24 +64,31 @@ export const externalAdsCostTile = (
let mathHogql: string
- if (tileColumnSelection === 'roas') {
- const costColumn = sanitizeColumnName(table.source_map.cost)
- const conversionValueColumn = table.source_map.reported_conversion_value
- ? sanitizeColumnName(table.source_map.reported_conversion_value)
- : null
- if (!costColumn || !conversionValueColumn) {
+ if (
+ tileColumnSelection === 'cpc' ||
+ tileColumnSelection === 'ctr' ||
+ tileColumnSelection === 'roas' ||
+ tileColumnSelection === 'cost_per_reported_conversion'
+ ) {
+ const ratioColumns = {
+ cpc: [MarketingAnalyticsColumnsSchemaNames.Cost, MarketingAnalyticsColumnsSchemaNames.Clicks],
+ ctr: [MarketingAnalyticsColumnsSchemaNames.Clicks, MarketingAnalyticsColumnsSchemaNames.Impressions],
+ roas: [
+ MarketingAnalyticsColumnsSchemaNames.ReportedConversionValue,
+ MarketingAnalyticsColumnsSchemaNames.Cost,
+ ],
+ cost_per_reported_conversion: [
+ MarketingAnalyticsColumnsSchemaNames.Cost,
+ MarketingAnalyticsColumnsSchemaNames.ReportedConversion,
+ ],
+ } as const
+ const [numeratorColumn, denominatorColumn] = ratioColumns[tileColumnSelection]
+ const numerator = externalAdsCostTile(table, baseCurrency, numeratorColumn)
+ const denominator = externalAdsCostTile(table, baseCurrency, denominatorColumn)
+ if (!numerator || !denominator) {
return null
}
- mathHogql = `${sumSafeFloat(conversionValueColumn)} / nullIf(SUM(toFloat(${costColumn})), 0)`
- } else if (tileColumnSelection === 'cost_per_reported_conversion') {
- const costColumn = sanitizeColumnName(table.source_map.cost)
- const conversionColumn = table.source_map.reported_conversion
- ? sanitizeColumnName(table.source_map.reported_conversion)
- : null
- if (!costColumn || !conversionColumn) {
- return null
- }
- mathHogql = `SUM(toFloat(${costColumn})) / nullIf(${sumSafeFloat(conversionColumn)}, 0)`
+ mathHogql = `(${numerator.math_hogql}) / nullIf(${denominator.math_hogql}, 0)${tileColumnSelection === 'ctr' ? ' * 100' : ''}`
} else {
const rawColumn = table.source_map[tileColumnSelection]
const column = rawColumn ? sanitizeColumnName(rawColumn) : null
@@ -86,7 +98,10 @@ export const externalAdsCostTile = (
const currencyExpr = buildCurrencyExpr(table.source_map.currency, baseCurrency)
const dateExpr = buildDateExpr(table)
const safeCurrency = sanitizeCurrencyCode(baseCurrency) ?? 'USD'
- mathHogql = `SUM(convertCurrency(${currencyExpr}, '${safeCurrency}', ${safeFloat(column)}, _toDate(${dateExpr})))`
+ mathHogql =
+ tileColumnSelection === 'cost' || tileColumnSelection === 'reported_conversion_value'
+ ? `SUM(convertCurrency(${currencyExpr}, '${safeCurrency}', ${safeFloat(column)}, _toDate(${dateExpr})))`
+ : sumSafeFloat(column)
}
const dateField = sanitizeColumnName(table.source_map.date)
diff --git a/frontend/src/scenes/web-analytics/tabs/marketing-analytics/frontend/logic/utils.test.ts b/frontend/src/scenes/web-analytics/tabs/marketing-analytics/frontend/logic/utils.test.ts
index c2e0b135e17d..8faa1e5c7c3c 100644
--- a/frontend/src/scenes/web-analytics/tabs/marketing-analytics/frontend/logic/utils.test.ts
+++ b/frontend/src/scenes/web-analytics/tabs/marketing-analytics/frontend/logic/utils.test.ts
@@ -595,6 +595,16 @@ describe('marketing analytics utils', () => {
expect(mathHogql).not.toContain(`${costColumn} / 1000000`)
})
+ it.each(['cpc', 'cost_per_reported_conversion', 'roas'] as const)(
+ 'converts Google Ads %s to the project currency',
+ (metric) => {
+ const source = makeMockSource('GoogleAds', sourceFields.GoogleAds)
+ expect(createMarketingTile(source, metric, 'EUR')?.math_hogql).toContain(
+ "convertCurrency(coalesce(customer_currency_code, 'EUR'), 'EUR'"
+ )
+ }
+ )
+
const testCases = VALID_NATIVE_MARKETING_SOURCES.flatMap((sourceType) =>
ALL_TILE_COLUMNS.map(
(column) =>
diff --git a/frontend/src/scenes/web-analytics/tabs/marketing-analytics/frontend/logic/utils.ts b/frontend/src/scenes/web-analytics/tabs/marketing-analytics/frontend/logic/utils.ts
index eee849e815ce..ce05b6cde1ee 100644
--- a/frontend/src/scenes/web-analytics/tabs/marketing-analytics/frontend/logic/utils.ts
+++ b/frontend/src/scenes/web-analytics/tabs/marketing-analytics/frontend/logic/utils.ts
@@ -420,9 +420,11 @@ export type validColumnsForTiles =
>
| 'roas'
| 'cost_per_reported_conversion'
+ | 'cpc'
+ | 'ctr'
// Raw column types that have actual column mappings (excludes calculated fields like ROAS)
-export type rawColumnsForTiles = Exclude<validColumnsForTiles, 'roas' | 'cost_per_reported_conversion'>
+export type rawColumnsForTiles = Exclude<validColumnsForTiles, 'roas' | 'cost_per_reported_conversion' | 'cpc' | 'ctr'>
interface ColumnConfig {
name: string
@@ -976,7 +978,8 @@ export function createMarketingTile(
tileColumnSelection === MarketingAnalyticsColumnsSchemaNames.Cost ||
tileColumnSelection === MarketingAnalyticsColumnsSchemaNames.ReportedConversionValue ||
tileColumnSelection === 'roas' ||
- tileColumnSelection === 'cost_per_reported_conversion'
+ tileColumnSelection === 'cost_per_reported_conversion' ||
+ tileColumnSelection === 'cpc'
if (monetaryColumn && !('currency_code' in table.fields)) {
return null
}
@@ -994,7 +997,8 @@ export function createMarketingTile(
tileColumnSelection === MarketingAnalyticsColumnsSchemaNames.Cost ||
tileColumnSelection === MarketingAnalyticsColumnsSchemaNames.ReportedConversionValue ||
tileColumnSelection === 'roas' ||
- tileColumnSelection === 'cost_per_reported_conversion'
+ tileColumnSelection === 'cost_per_reported_conversion' ||
+ tileColumnSelection === 'cpc'
const currencyColumn = sourceType === 'AmazonAds' ? 'campaign_budget_currency_code' : 'currency'
if (monetaryColumn && !(currencyColumn in table.fields)) {
return null
@@ -1005,7 +1009,10 @@ export function createMarketingTile(
if (!['campaign_id', 'start_time', 'impressions', 'clicks', 'spend'].every((field) => field in table.fields)) {
return null
}
- if (tileColumnSelection === MarketingAnalyticsColumnsSchemaNames.Cost && !('currency_code' in table.fields)) {
+ if (
+ (tileColumnSelection === MarketingAnalyticsColumnsSchemaNames.Cost || tileColumnSelection === 'cpc') &&
+ !('currency_code' in table.fields)
+ ) {
return null
}
}
@@ -1019,6 +1026,15 @@ export function createMarketingTile(
}
: tileConfig.columnMappings
+ if (tileColumnSelection === 'cpc' || tileColumnSelection === 'ctr') {
+ const clicks = sumSafeFloat(tileConfig.columnMappings.clicks)
+ const expression =
+ tileColumnSelection === 'cpc'
+ ? `${wrapWithCurrencyConversion(buildNativeCostExpr(tileConfig.columnMappings), tileConfig.columnMappings, table, baseCurrency)} / nullIf(${clicks}, 0)`
+ : `${clicks} / nullIf(${sumSafeFloat(tileConfig.columnMappings.impressions)}, 0) * 100`
+ return buildNativeTileNode(table, integrationConfig, tileConfig, tileColumnSelection, expression)
+ }
+
// Handle ROAS (Return on Ad Spend) - calculated as conversion_value / cost
if (tileColumnSelection === 'roas') {
const costExpr = buildNativeCostExpr(tileConfig.columnMappings)
@@ -1030,7 +1046,7 @@ export function createMarketingTile(
)
let totalValueExpr = conversionValueExpr
let totalCostExpr = `SUM(${costExpr})`
- if (tileConfig.columnMappings.currencyTimestampColumn && conversionValueExpr !== '0') {
+ if (conversionValueExpr !== '0') {
const perRowValue =
tileConfig.specialConversionLogic?.(table, MarketingAnalyticsColumnsSchemaNames.ReportedConversionValue)
?.perRowValueExpr ?? safeFloat(tileConfig.columnMappings.reportedConversionValue)
@@ -1050,9 +1066,7 @@ export function createMarketingTile(
MarketingAnalyticsColumnsSchemaNames.ReportedConversion,
tileConfig.columnMappings.reportedConversion
)
- const totalCostExpr = tileConfig.columnMappings.currencyTimestampColumn
- ? wrapWithCurrencyConversion(costExpr, tileConfig.columnMappings, table, baseCurrency)
- : `SUM(${costExpr})`
+ const totalCostExpr = wrapWithCurrencyConversion(costExpr, tileConfig.columnMappings, table, baseCurrency)
const mathHogql = conversionExpr === '0' ? '0' : `${totalCostExpr} / nullIf(${conversionExpr}, 0)`
return buildNativeTileNode(table, integrationConfig, tileConfig, tileColumnSelection, mathHogql)
}
diff --git a/frontend/src/scenes/web-analytics/tiles/WebAnalyticsTile.tsx b/frontend/src/scenes/web-analytics/tiles/WebAnalyticsTile.tsx
index 5ec7491160f5..5936f0c56a2c 100644
--- a/frontend/src/scenes/web-analytics/tiles/WebAnalyticsTile.tsx
+++ b/frontend/src/scenes/web-analytics/tiles/WebAnalyticsTile.tsx
@@ -991,6 +991,8 @@ export const MarketingAnalyticsTrendTile = ({
value: MarketingAnalyticsColumnsSchemaNames.ReportedConversionValue,
label: 'Reported conversion value',
},
+ { value: 'cpc', label: 'CPC' },
+ { value: 'ctr', label: 'CTR' },
{ value: 'roas', label: 'Reported ROAS' },
{ value: 'cost_per_reported_conversion', label: 'Cost per reported conversion' },
]
diff --git a/products/marketing_analytics/frontend/AdPerformanceOverview.stories.tsx b/products/marketing_analytics/frontend/AdPerformanceOverview.stories.tsx
new file mode 100644
index 000000000000..789d8a3a0903
--- /dev/null
+++ b/products/marketing_analytics/frontend/AdPerformanceOverview.stories.tsx
@@ -0,0 +1,212 @@
+import { Meta, StoryObj } from '@storybook/react'
+import { useActions, useValues } from 'kea'
+import { useEffect } from 'react'
+
+import { FEATURE_FLAGS } from 'lib/constants'
+import { databaseTableListLogic } from 'scenes/data-management/database/databaseTableListLogic'
+import { MarketingAnalyticsOverview } from 'scenes/web-analytics/tabs/marketing-analytics/frontend/components/MarketingAnalyticsOverview/MarketingAnalyticsOverview'
+import { marketingAnalyticsLogic } from 'scenes/web-analytics/tabs/marketing-analytics/frontend/logic/marketingAnalyticsLogic'
+import { marketingAnalyticsTilesLogic } from 'scenes/web-analytics/tabs/marketing-analytics/frontend/logic/marketingAnalyticsTilesLogic'
+import { MarketingAnalyticsTrendTile } from 'scenes/web-analytics/tiles/WebAnalyticsTile'
+
+import { useStorybookMocks } from '~/mocks/browser'
+import {
+ DatabaseSchemaDataWarehouseTable,
+ InsightVizNode,
+ MARKETING_INTEGRATION_CONFIGS,
+ MarketingAnalyticsAggregatedQueryResponse,
+ NodeKind,
+ TrendsQuery,
+} from '~/queries/schema/schema-general'
+import { ExternalDataSource } from '~/types'
+
+const days = ['2026-09-10', '2026-09-11', '2026-09-12', '2026-09-13', '2026-09-14', '2026-09-15', '2026-09-16']
+const data: Record<string, number[]> = {
+ cost: [2200, 2600, 2100, 3100, 3000, 2400, 2800],
+ clicks: [1200, 1400, 1000, 1800, 1500, 1300, 1400],
+ impressions: [80000, 90000, 70000, 110000, 100000, 90000, 100000],
+ reported_conversion: [40, 50, 30, 70, 60, 45, 55],
+ reported_conversion_value: [11000, 13000, 9000, 16000, 15000, 12000, 15000],
+}
+for (const [metric, numerator, denominator, scale] of [
+ ['cpc', 'cost', 'clicks', 1],
+ ['ctr', 'clicks', 'impressions', 100],
+ ['roas', 'reported_conversion_value', 'cost', 1],
+ ['cost_per_reported_conversion', 'cost', 'reported_conversion', 1],
+] as const) {
+ data[metric] = data[numerator].map((value, index) => (value / data[denominator][index]) * scale)
+}
+
+const schemaNames = [
+ MARKETING_INTEGRATION_CONFIGS.GoogleAds.campaignTableName,
+ MARKETING_INTEGRATION_CONFIGS.GoogleAds.statsTableName,
+]
+const source = {
+ id: 'example-google-ads',
+ source_type: 'GoogleAds',
+ prefix: 'example',
+ status: 'Completed',
+ schemas: schemaNames.map((name) => ({ id: `example-${name}`, name, should_sync: true, status: 'Completed' })),
+} as ExternalDataSource
+const tables = Object.fromEntries(
+ schemaNames.map((name) => [
+ name,
+ {
+ id: `table-${name}`,
+ name,
+ type: 'data_warehouse',
+ schema: { id: `example-${name}`, name, should_sync: true, incremental: false },
+ fields: Object.fromEntries(
+ [
+ 'campaign_id',
+ 'id',
+ 'name',
+ 'segments_date',
+ 'customer_currency_code',
+ 'metrics_cost_micros',
+ 'metrics_clicks',
+ 'metrics_impressions',
+ 'metrics_conversions',
+ 'metrics_conversions_value',
+ ].map((field) => [field, { name: field, hogql_value: field, type: 'float', schema_valid: true }])
+ ),
+ } satisfies DatabaseSchemaDataWarehouseTable,
+ ])
+)
+
+const summary: MarketingAnalyticsAggregatedQueryResponse = {
+ results: {
+ Cost: {
+ key: 'Cost',
+ value: 18200,
+ previous: 20000,
+ changeFromPreviousPct: -9,
+ kind: 'currency',
+ isIncreaseBad: true,
+ },
+ Clicks: { key: 'Clicks', value: 9600, previous: 8000, changeFromPreviousPct: 20, kind: 'unit' },
+ Impressions: { key: 'Impressions', value: 640000, previous: 600000, changeFromPreviousPct: 7, kind: 'unit' },
+ 'Reported Conversions': {
+ key: 'Reported Conversions',
+ value: 350,
+ previous: 300,
+ changeFromPreviousPct: 17,
+ kind: 'unit',
+ },
+ 'Reported Conversion Value': {
+ key: 'Reported Conversion Value',
+ value: 91000,
+ previous: 75000,
+ changeFromPreviousPct: 21,
+ kind: 'currency',
+ },
+ 'Trial started': { key: 'Trial started', value: 620, previous: 540, changeFromPreviousPct: 15, kind: 'unit' },
+ 'Demo booked': { key: 'Demo booked', value: 84, previous: 70, changeFromPreviousPct: 20, kind: 'unit' },
+ CPC: {
+ key: 'CPC',
+ value: 1.9,
+ previous: 2.5,
+ changeFromPreviousPct: -24,
+ kind: 'currency',
+ isIncreaseBad: true,
+ },
+ CTR: { key: 'CTR', value: 1.5, previous: 1.33, changeFromPreviousPct: 13, kind: 'percentage' },
+ 'Cost per Reported Conversions': {
+ key: 'Cost per Reported Conversions',
+ value: 52,
+ previous: 66.67,
+ changeFromPreviousPct: -22,
+ kind: 'currency',
+ isIncreaseBad: true,
+ },
+ 'Reported ROAS': { key: 'Reported ROAS', value: 5, previous: 3.75, changeFromPreviousPct: 33, kind: 'unit' },
+ 'Cost per Trial started': { key: 'Cost per Trial started', value: 29.35, kind: 'currency' },
+ 'Cost per Demo booked': { key: 'Cost per Demo booked', value: 216.67, kind: 'currency' },
+ },
+}
+
+const meta: Meta<typeof MarketingAnalyticsOverview> = {
+ title: 'Marketing Analytics/Ad performance metric trends',
+ component: MarketingAnalyticsOverview,
+ parameters: { layout: 'padded', mockDate: '2026-09-16', featureFlags: [FEATURE_FLAGS.WEB_ANALYTICS_MARKETING] },
+ args: {
+ query: {
+ kind: NodeKind.MarketingAnalyticsAggregatedQuery,
+ properties: [],
+ dateRange: { date_from: '2026-09-10', date_to: '2026-09-16' },
+ },
+ cachedResults: summary,
+ context: {},
+ uniqueKey: 'ad-performance-trends-story',
+ },
+}
+export default meta
+type Story = StoryObj<typeof MarketingAnalyticsOverview>
+
+export const WithTrends: Story = {
+ render: function Render(args) {
+ const { loadSourcesSuccess, setDatesAndInterval } = useActions(marketingAnalyticsLogic)
+ const { marketingChartTile } = useValues(marketingAnalyticsTilesLogic)
+ useStorybookMocks({
+ get: {
+ '/api/projects/:team_id/external_data_sources/wizard/': {},
+ '/api/environments/:team_id/external_data_sources/': {
+ count: 1,
+ results: [source],
+ next: null,
+ previous: null,
+ },
+ '/api/projects/:team_id/marketing_analytics/source_validation/': { errors_by_source: {} },
+ },
+ post: {
+ '/api/environments/:team_id/query/:kind/': async ({ request }) => {
+ const { query } = (await request.json()) as { query: TrendsQuery }
+ if (query.kind === (NodeKind.DatabaseSchemaQuery as string)) {
+ return [200, { tables, joins: [] }]
+ }
+ return [
+ 200,
+ {
+ results: (query.series ?? []).map((series, index) => {
+ const metric =
+ series.custom_name && data[series.custom_name]
+ ? series.custom_name
+ : marketingAnalyticsLogic.values.tileColumnSelection
+ return {
+ action: { ...series, type: 'data_warehouse', order: index },
+ order: index,
+ label: series.custom_name ?? metric,
+ days,
+ dates: days,
+ labels: days,
+ data: data[metric] ?? data.cost,
+ count: (data[metric] ?? data.cost).reduce((total, value) => total + value, 0),
+ aggregated_value: (data[metric] ?? data.cost).reduce(
+ (total, value) => total + value,
+ 0
+ ),
+ }
+ }),
+ },
+ ]
+ },
+ },
+ })
+ useEffect(() => {
+ loadSourcesSuccess({ count: 1, results: [source], next: null, previous: null })
+ databaseTableListLogic.actions.loadDatabaseSuccess({ tables, joins: [] })
+ setDatesAndInterval('2026-09-10', '2026-09-16', 'day')
+ }, [loadSourcesSuccess, setDatesAndInterval])
+ return (
+ <div className="flex flex-col gap-4">
+ <MarketingAnalyticsOverview {...args} />
+ <MarketingAnalyticsTrendTile
+ query={marketingChartTile.query as InsightVizNode}
+ insightProps={marketingChartTile.insightProps}
+ uniqueKey="ad-performance-main-trend-story"
+ showIntervalTile
+ />
+ </div>
+ )
+ },
+}
diff --git a/products/marketing_analytics/frontend/adPerformanceMetrics.test.ts b/products/marketing_analytics/frontend/adPerformanceMetrics.test.ts
new file mode 100644
index 000000000000..36b7dd22d547
--- /dev/null
+++ b/products/marketing_analytics/frontend/adPerformanceMetrics.test.ts
@@ -0,0 +1,144 @@
+import type { ExternalTable } from 'scenes/web-analytics/tabs/marketing-analytics/frontend/logic/marketingAnalyticsLogic'
+import { externalAdsCostTile } from 'scenes/web-analytics/tabs/marketing-analytics/frontend/logic/marketingCostTile'
+
+import { MarketingAnalyticsColumnsSchemaNames, NodeKind, SourceMap } from '~/queries/schema/schema-general'
+import { DataWarehouseSettingsTab, PropertyFilterType, PropertyOperator, TrendResult } from '~/types'
+
+import { adPerformanceSparklines, buildAdPerformanceTrendsQuery } from './adPerformanceMetrics'
+
+const row = (metric: string, data: number[], days = ['2026-01-01', '2026-01-02']): TrendResult =>
+ ({
+ action: { custom_name: metric },
+ data,
+ days,
+ labels: days,
+ label: metric,
+ count: 0,
+ aggregated_value: 0,
+ }) as TrendResult
+
+const table: ExternalTable = {
+ id: 'table-one',
+ source_map_id: 'source-one',
+ name: 'example_ad_stats',
+ source_type: 'manual',
+ source_prefix: '',
+ columns: [],
+ url_pattern: '',
+ sourceUrl: '',
+ external_type: DataWarehouseSettingsTab.Managed,
+ schema_name: 'example_ad_stats',
+ dw_source_type: 'manual',
+ source_map: {
+ ...(Object.fromEntries(
+ Object.values(MarketingAnalyticsColumnsSchemaNames).map((key) => [key, undefined])
+ ) as SourceMap),
+ date: 'day',
+ cost: 'spend',
+ clicks: 'clicks',
+ impressions: 'impressions',
+ reported_conversion: 'conversions',
+ reported_conversion_value: 'revenue',
+ currency: 'const:EUR',
+ },
+}
+
+describe('Ad performance sparklines', () => {
+ it('sums sources by date before calculating ratios and leaves zero denominators as gaps', () => {
+ const metrics = adPerformanceSparklines([
+ row('cost', [10, 20]),
+ row('cost', [90, 30]),
+ row('clicks', [5, 0]),
+ row('clicks', [15, 0]),
+ row('impressions', [200, 100]),
+ row('reported_conversion', [2, 0]),
+ row('reported_conversion_value', [300, 0]),
+ ])
+ expect(metrics.cost?.data).toEqual([100, 50])
+ expect(metrics.cpc?.data).toEqual([5, NaN])
+ expect(metrics.ctr?.data).toEqual([10, 0])
+ expect(metrics.roas?.data).toEqual([3, 0])
+ expect(metrics.cost_per_reported_conversion?.data).toEqual([50, NaN])
+ })
+
+ it('aligns dates, preserves unknown values, and excludes attributed goals and comparison periods', () => {
+ const metrics = adPerformanceSparklines([
+ row('cost', [20, 10], ['2026-01-02', '2026-01-01']),
+ row('clicks', [5, NaN]),
+ row('trial started', [100, 200]),
+ { ...row('cost', [500, 600]), compare: true },
+ ])
+ expect(metrics.cost).toEqual({ labels: ['2026-01-01', '2026-01-02'], data: [10, 20] })
+ expect(metrics.cpc?.data).toEqual([2, NaN])
+ expect(Object.keys(metrics)).not.toContain('trial started')
+ })
+
+ it.each([false, true])('keeps request filters and current-period dates with precompute=%s', (precomputed) => {
+ const query = buildAdPerformanceTrendsQuery(
+ {
+ kind: NodeKind.MarketingAnalyticsAggregatedQuery,
+ properties: [],
+ dateRange: { date_from: '-30d', date_to: null },
+ compareFilter: { compare: true },
+ integrationFilter: { integrationSourceIds: ['source-one'] },
+ },
+ 'day',
+ [],
+ [table, { ...table, source_map_id: 'source-two' }],
+ 'USD',
+ precomputed
+ )
+ expect(query.series).toHaveLength(5)
+ expect(query.dateRange).toEqual({ date_from: '-30d', date_to: null })
+ expect(query.compareFilter).toBeUndefined()
+ if (precomputed) {
+ expect(query.series[0].properties).toContainEqual({
+ type: PropertyFilterType.DataWarehouse,
+ key: 'source_id',
+ operator: PropertyOperator.Exact,
+ value: ['source-one'],
+ })
+ } else {
+ expect(query.series.every((series) => 'table_name' in series && series.table_name === table.name)).toBe(
+ true
+ )
+ expect((query.series[0] as { math_hogql: string }).math_hogql).toContain('convertCurrency')
+ expect((query.series[1] as { math_hogql: string }).math_hogql).not.toContain('convertCurrency')
+ expect((query.series[3] as { math_hogql: string }).math_hogql).not.toContain('convertCurrency')
+ }
+ })
+
+ it.each([
+ [
+ 'cost_per_reported_conversion',
+ MarketingAnalyticsColumnsSchemaNames.Cost,
+ MarketingAnalyticsColumnsSchemaNames.ReportedConversion,
+ ],
+ [
+ 'roas',
+ MarketingAnalyticsColumnsSchemaNames.ReportedConversionValue,
+ MarketingAnalyticsColumnsSchemaNames.Cost,
+ ],
+ ] as const)(
+ 'keeps the external %s trend in the same currency as its sparkline',
+ (metric, numerator, denominator) => {
+ const top = externalAdsCostTile(table, 'USD', numerator)?.math_hogql
+ const bottom = externalAdsCostTile(table, 'USD', denominator)?.math_hogql
+ expect(externalAdsCostTile(table, 'USD', metric)?.math_hogql).toBe(`(${top}) / nullIf(${bottom}, 0)`)
+ expect(externalAdsCostTile(table, 'USD', metric)?.math_hogql).toContain("convertCurrency('EUR', 'USD'")
+ }
+ )
+
+ it('does not request data when there are no supported sources', () => {
+ expect(
+ buildAdPerformanceTrendsQuery(
+ { kind: NodeKind.MarketingAnalyticsAggregatedQuery, properties: [] },
+ 'day',
+ [],
+ [],
+ 'USD',
+ false
+ ).series
+ ).toEqual([])
+ })
+})
diff --git a/products/marketing_analytics/frontend/adPerformanceMetrics.ts b/products/marketing_analytics/frontend/adPerformanceMetrics.ts
new file mode 100644
index 000000000000..770e3ee078c4
--- /dev/null
+++ b/products/marketing_analytics/frontend/adPerformanceMetrics.ts
@@ -0,0 +1,187 @@
+import type {
+ ExternalTable,
+ NativeSource,
+} from 'scenes/web-analytics/tabs/marketing-analytics/frontend/logic/marketingAnalyticsLogic'
+import { externalAdsCostTile } from 'scenes/web-analytics/tabs/marketing-analytics/frontend/logic/marketingCostTile'
+import {
+ MarketingDashboardMapper,
+ type validColumnsForTiles,
+} from 'scenes/web-analytics/tabs/marketing-analytics/frontend/logic/utils'
+
+import {
+ DataWarehouseNode,
+ MarketingAnalyticsAggregatedQuery,
+ MarketingAnalyticsBaseColumns,
+ MarketingAnalyticsColumnsSchemaNames,
+ NodeKind,
+ TrendsQuery,
+} from '~/queries/schema/schema-general'
+import {
+ ChartDisplayType,
+ HogQLMathType,
+ IntervalType,
+ PropertyFilterType,
+ PropertyOperator,
+ TrendResult,
+} from '~/types'
+
+export const AD_PERFORMANCE_METRICS: Record<string, validColumnsForTiles> = {
+ [MarketingAnalyticsBaseColumns.Cost]: MarketingAnalyticsColumnsSchemaNames.Cost,
+ [MarketingAnalyticsBaseColumns.Clicks]: MarketingAnalyticsColumnsSchemaNames.Clicks,
+ [MarketingAnalyticsBaseColumns.Impressions]: MarketingAnalyticsColumnsSchemaNames.Impressions,
+ [MarketingAnalyticsBaseColumns.ReportedConversions]: MarketingAnalyticsColumnsSchemaNames.ReportedConversion,
+ [MarketingAnalyticsBaseColumns.ReportedConversionValue]:
+ MarketingAnalyticsColumnsSchemaNames.ReportedConversionValue,
+ [MarketingAnalyticsBaseColumns.CPC]: 'cpc',
+ [MarketingAnalyticsBaseColumns.CTR]: 'ctr',
+ [MarketingAnalyticsBaseColumns.ReportedROAS]: 'roas',
+ [MarketingAnalyticsBaseColumns.CostPerReportedConversions]: 'cost_per_reported_conversion',
+}
+
+export function adPerformanceMetricKey(key: string): validColumnsForTiles | undefined {
+ return Object.hasOwn(AD_PERFORMANCE_METRICS, key) ? AD_PERFORMANCE_METRICS[key] : undefined
+}
+
+const RAW_METRICS = [
+ MarketingAnalyticsColumnsSchemaNames.Cost,
+ MarketingAnalyticsColumnsSchemaNames.Clicks,
+ MarketingAnalyticsColumnsSchemaNames.Impressions,
+ MarketingAnalyticsColumnsSchemaNames.ReportedConversion,
+ MarketingAnalyticsColumnsSchemaNames.ReportedConversionValue,
+] as const
+const PRECOMPUTED_FIELDS = [
+ 'cost',
+ 'clicks',
+ 'impressions',
+ 'reported_conversions',
+ 'reported_conversion_value',
+] as const
+
+export interface AdPerformanceSparkline {
+ data: number[]
+ labels: string[]
+}
+
+export function buildAdPerformanceTrendsQuery(
+ query: MarketingAnalyticsAggregatedQuery,
+ interval: IntervalType,
+ nativeSources: NativeSource[],
+ externalTables: ExternalTable[],
+ baseCurrency: string,
+ precomputed: boolean
+): TrendsQuery {
+ const selectedIds = query.integrationFilter?.integrationSourceIds ?? []
+ const native = nativeSources.filter((source) => !selectedIds.length || selectedIds.includes(source.source.id))
+ const external = externalTables.filter((table) => !selectedIds.length || selectedIds.includes(table.source_map_id))
+ const series: DataWarehouseNode[] = []
+ if (native.length || external.length) {
+ RAW_METRICS.forEach((metric, index) => {
+ const nodes: DataWarehouseNode[] = precomputed
+ ? [
+ {
+ kind: NodeKind.DataWarehouseNode,
+ id: 'marketing_costs_precomputed',
+ table_name: 'marketing_costs_precomputed',
+ id_field: 'campaign_id',
+ distinct_id_field: 'campaign_id',
+ timestamp_field: 'timestamp',
+ math: HogQLMathType.HogQL,
+ math_hogql: `sum(${PRECOMPUTED_FIELDS[index]})`,
+ properties: [
+ { type: PropertyFilterType.HogQL, key: "grain = 'campaign' AND expires_at > today()" },
+ ...(selectedIds.length
+ ? [
+ {
+ type: PropertyFilterType.DataWarehouse as const,
+ key: 'source_id',
+ operator: PropertyOperator.Exact,
+ value: selectedIds,
+ },
+ ]
+ : []),
+ ],
+ },
+ ]
+ : [
+ ...native.map((source) => MarketingDashboardMapper(source, metric, baseCurrency)),
+ ...external.map((table) => externalAdsCostTile(table, baseCurrency, metric)),
+ ].filter((node): node is DataWarehouseNode => node !== null)
+ series.push(...nodes.map((node) => ({ ...node, custom_name: metric })))
+ })
+ }
+ return {
+ kind: NodeKind.TrendsQuery,
+ series,
+ interval,
+ dateRange: query.dateRange,
+ properties: query.properties,
+ filterTestAccounts: query.filterTestAccounts,
+ trendsFilter: { display: ChartDisplayType.ActionsLineGraph },
+ tags: query.tags,
+ }
+}
+
+export function adPerformanceSparklines(
+ results: TrendResult[]
+): Partial<Record<validColumnsForTiles, AdPerformanceSparkline>> {
+ const buckets = new Map<string, Map<string, number>>()
+ for (const row of results) {
+ const metric = row.action?.custom_name
+ if (row.compare || !metric || !RAW_METRICS.some((key) => key === metric)) {
+ continue
+ }
+ const values = buckets.get(metric) ?? new Map<string, number>()
+ row.days.forEach((day, index) => {
+ const value = row.data[index]
+ values.set(
+ day,
+ (values.get(day) ?? 0) + (typeof value === 'number' && Number.isFinite(value) ? value : NaN)
+ )
+ })
+ buckets.set(metric, values)
+ }
+ const labels = [...new Set([...buckets.values()].flatMap((values) => [...values.keys()]))].sort()
+ const output: Partial<Record<validColumnsForTiles, AdPerformanceSparkline>> = {}
+ const add = (metric: validColumnsForTiles, data: number[]): void => {
+ if (data.some(Number.isFinite)) {
+ output[metric] = { data, labels }
+ }
+ }
+ for (const metric of RAW_METRICS) {
+ const values = buckets.get(metric)
+ if (values) {
+ add(
+ metric,
+ labels.map((label) => values.get(label) ?? NaN)
+ )
+ }
+ }
+ const ratio = (
+ metric: validColumnsForTiles,
+ numerator: (typeof RAW_METRICS)[number],
+ denominator: (typeof RAW_METRICS)[number],
+ scale = 1
+ ): void => {
+ const top = output[numerator]?.data
+ const bottom = output[denominator]?.data
+ if (top && bottom) {
+ add(
+ metric,
+ top.map((value, index) => (bottom[index] ? (value / bottom[index]) * scale : NaN))
+ )
+ }
+ }
+ ratio('cpc', MarketingAnalyticsColumnsSchemaNames.Cost, MarketingAnalyticsColumnsSchemaNames.Clicks)
+ ratio('ctr', MarketingAnalyticsColumnsSchemaNames.Clicks, MarketingAnalyticsColumnsSchemaNames.Impressions, 100)
+ ratio(
+ 'roas',
+ MarketingAnalyticsColumnsSchemaNames.ReportedConversionValue,
+ MarketingAnalyticsColumnsSchemaNames.Cost
+ )
+ ratio(
+ 'cost_per_reported_conversion',
+ MarketingAnalyticsColumnsSchemaNames.Cost,
+ MarketingAnalyticsColumnsSchemaNames.ReportedConversion
+ )
+ return output
+}
diff --git a/products/marketing_analytics/frontend/adPerformanceTrendsLogic.ts b/products/marketing_analytics/frontend/adPerformanceTrendsLogic.ts
new file mode 100644
index 000000000000..c6bb4223ec1a
--- /dev/null
+++ b/products/marketing_analytics/frontend/adPerformanceTrendsLogic.ts
@@ -0,0 +1,65 @@
+import { LogicWrapper, MakeLogicType, connect, kea, key, path, props, propsChanged, selectors } from 'kea'
+
+import type { validColumnsForTiles } from 'scenes/web-analytics/tabs/marketing-analytics/frontend/logic/utils'
+
+import { DataNodeLogicProps, dataNodeLogic, dataNodeLogicType } from '~/queries/nodes/DataNode/dataNodeLogic'
+import { TrendsQuery, TrendsQueryResponse } from '~/queries/schema/schema-general'
+import { TrendResult } from '~/types'
+
+import { AdPerformanceSparkline, adPerformanceSparklines } from './adPerformanceMetrics'
+
+export interface AdPerformanceTrendsLogicProps {
+ query: TrendsQuery
+ queryKey: string
+ dataNodeCollectionId: string
+}
+
+function dataProps(props: AdPerformanceTrendsLogicProps): DataNodeLogicProps {
+ return {
+ key: props.queryKey,
+ query: props.query,
+ doNotLoad: !props.query.series.length,
+ dataNodeCollectionId: props.dataNodeCollectionId,
+ }
+}
+
+export interface adPerformanceTrendsLogicValues extends Pick<
+ dataNodeLogicType['values'],
+ 'response' | 'responseLoading' | 'responseError'
+> {
+ sparklines: Partial<Record<validColumnsForTiles, AdPerformanceSparkline>>
+}
+
+export type adPerformanceTrendsLogicType = MakeLogicType<
+ adPerformanceTrendsLogicValues,
+ {},
+ AdPerformanceTrendsLogicProps
+>
+
+export const adPerformanceTrendsLogic: LogicWrapper<adPerformanceTrendsLogicType> = kea<adPerformanceTrendsLogicType>([
+ props({} as AdPerformanceTrendsLogicProps),
+ key(({ queryKey }) => queryKey),
+ path((key) => ['products', 'marketingAnalytics', 'adPerformanceTrendsLogic', key]),
+ connect((props: AdPerformanceTrendsLogicProps) => ({
+ values: [dataNodeLogic(dataProps(props)), ['response', 'responseLoading', 'responseError']],
+ })),
+ selectors({
+ sparklines: [
+ (s) => [s.response, s.responseLoading, s.responseError],
+ (response, loading, error): Partial<Record<validColumnsForTiles, AdPerformanceSparkline>> =>
+ loading || error
+ ? {}
+ : adPerformanceSparklines(
+ ((response as TrendsQueryResponse | null)?.results ?? []) as TrendResult[]
+ ),
+ ],
+ }),
+ propsChanged(({ props }, oldProps) => {
+ if (JSON.stringify(props.query) !== JSON.stringify(oldProps.query)) {
+ const logic = dataNodeLogic.findMounted({ key: props.queryKey })
+ logic?.actions.abortAnyRunningQuery()
+ logic?.actions.clearResponse()
+ dataNodeLogic(dataProps(props))
+ }
+ }),
+])