Registered
936
754 external · 182 Bentley
AI Pillar Overview
Campaign performance reporting with traceable KPIs, sourced from the AI Pillar data layer
SAMPLE DATA · NOT LIVE
Executive Performance Eight headline KPIs · click a metric for its lineage
Data note: All displayed performance figures and comparisons are illustrative; actual production values require validated source connections.
Campaign Performance Trend
Weekly engaged known contacts and campaign MQLs · source-normalized and deduplicated
Engaged known contactsCampaign MQLs
Audience Journey
Distinct identified people, then qualified downstream stages
The progression numerator requires a later action after the initial qualified engagement.
Engaged Audience by Channel
Unique known contacts, assigned to a first qualifying engagement source
Visibility vs. Response
Different units shown separately — not summed
Leadership Interpretation
What the executive page is designed to answer
1 · Are we visible?Reach, SOV, discoverability
2 · Are we resonating?Engaged contacts, actions
3 · Are we advancing?Progression, MQLs
4 · Is it translating to demand?Intent, influenced pipeline
eStore orders and revenue remain separate commercial outcomes rather than being added to qualified intent.
Event impact · who was in the room from Dallas cohort analysis
Turnout, the accounts and ARR in the room, and the pipeline that was already open at those accounts on the day · one row per AI Pillar event
| Event | Registered | Attended | External show rate | Accounts in the room | ARR in the room | Open pipeline in the room illustrative | Pipeline created ≤90 days after illustrative |
|---|---|---|---|---|---|---|---|
| Illuminate Dallas 2026 | 936 | 570 | 54% | 140 of 202 | $302M | $48.6M · 212 opps | $6.1M · 37 opps |
| Next AI Pillar events | Populate automatically once each event campaign sits under an AI Pillar sub-theme in Salesforce | ||||||
Open pipeline in the room = opportunities at attending accounts (ultimate parent) created before the event date and still open on it. Event Performance →
Visibility & Reach People-based reach separated from exposure and traffic
Unique Identifiable Reach by Channel
De-duplicated contact IDs; excludes anonymous pageviews and impressions
Share of Voice Trend
Bentley AI-term mentions / Bentley plus monitored competitor mentions
Web Discoverability
GA4 AI Hub and AI blog activity (not unique people)
LLM Discoverability
AI mentions, citations and referrals as distinct components
PR / Analyst Context
Visibility diagnostics only; never part of unique-person reach
Content & Audience Engagement
Content Engagement Actions by Channel
Interaction count, including anonymous social interactions · separate from known contact counts
Qualified Engagement Over Time
Content actions versus distinct engaged contacts (two different measures)
Content / Experience Performance
Illustrative assets and event themes across linked touchpoints
Email & Event Diagnostics
Operational metrics feeding the journey without inflating the executive reach count
Audience Progression & Marketing Qualification
Reached → Engaged → Progressed → MQL
Contact-level sequence; MQLs deduplicated across Salesforce AI Pillar subcampaigns
Progression by First Engagement Source
Progressed contacts / engaged contacts in each source
Journey Conversion
Conversion rates are people-based, not counts of repeated activities
Required Business Rules
Proposed data-layer logic applied in this demo
Identity resolutionContact/Lead master ID
Progression sequenceLater action only
Attribution window (demo)90 days
MQL program roll-upAI Pillar parent campaign
Internal AAG contactsExcluded from external MQL
Commercial Impact
Qualified Intent by Request Type
Distinct people in the headline KPI; requests can appear in multiple diagnostic types
Influenced Pipeline by Opportunity Stage
One opportunity counted once after an eligible prior campaign touch
Example Salesforce Opportunities
Synthetic opportunity rows and campaign-influence links; dollar values are not actual pipeline
eStore / Completed Commercial Results
Separate from requests, campaign influence, and pipeline
Influenced pipeline is a strategic relationship indicator, not proof that the campaign solely created the opportunity.
Event Performance Who registered, who showed up, which accounts and how much ARR and open pipeline were in the room
Attended
570
61% overall show rate
External show rate
54%
408 of 754 · ref: Sydney 53%, Berlin 56%
Bentley show rate
89%
162 of 182 · kept separate from external
Accounts in the room
140
of 202 registered companies (69%)
ARR in the room
$302M
global relationship ARR, existing customers
Open pipeline in the roomillustrative
$48.6M
212 open opps at attending accounts on the day
Pipeline created ≤90 days afterillustrative
$6.1M
37 new opps at attending accounts
Who showed up, and who didn’t from Dallas cohort analysis
Registered = attended + no-show · Bentley staff reported apart from external
| Cohort | Registered | Attended | No-show | Show rate | Share of no-shows |
|---|---|---|---|---|---|
| External | 754 | 408 | 346 | 54% | 95% |
| Bentley | 182 | 162 | 20 | 89% | 5% |
| Total | 936 | 570 | 366 | 61% | 100% |
External
408 attended346 no-show
Bentley
162 attended20
What drives no-shows? from Dallas cohort analysis
External registrants · show rate by registrant profile (average 54%; orange = below average)
| Registrant profile | Registered | Show rate |
|---|---|---|
| Based outside the US | 72 | |
| Registered with a personal email | 82 | |
| Account unknown or <$50K ARR | 273 | |
| Small segment account | 59 | |
| Only person from their company | 173 | |
| Lead (new prospect), not Contact | 283 | |
| Company sent a team of 2–3 | 184 | |
| Account with $250K–$1M ARR | 169 |
So whatCompanies sending 2–3 people and US registrants showed best; international, personal-email and solo registrants mostly didn’t. Plan on a ~53% external show rate and have account owners reconfirm priority registrants two weeks out.
Which kinds of accounts were in the room from Dallas cohort analysis
Industry derived from account name · 97% US · 75% Enterprise · 69% existing customer contacts
| Industry | Registered | Attended | Show rate |
|---|---|---|---|
| Engineering & design firms | 466 | 271 | |
| Construction & contractors | 52 | 32 | |
| Technology, industrial & other | 67 | 31 | |
| Utilities & energy | 30 | 20 | |
| Government & public agencies | 26 | 14 | |
| Rail, transit & ports | 16 | 12 | |
| Education | 15 | 9 | |
| Unknown (personal email) | 82 | 19 |
Account turnout and ARR from Dallas cohort analysis
Accounts grouped by how many of their registrants came · ARR on the ultimate-parent account
| Account turnout | Accounts | Attended | ARR (est.) | Share of ARR |
|---|---|---|---|---|
| Full: everyone came | 83 | 153 of 153 | $89.8M | 25% |
| Partial: some no-shows | 57 | 189 of 338 | $212.0M | 60% |
| None: no one came | 62 | 0 of 78 | $51.7M | 15% |
| Total registered | 202 | 342 of 569 | $353.6M | 100% |
So whatPartial-turnout accounts hold most of the ARR (57 accounts, ~$212M, 149 no-shows). Biggest gaps: WSP 2 of 17, HDR 3 of 11, CDM Smith 4 of 11, AECOM 6 of 15. The opportunity is lifting attendance inside high-value Enterprise accounts, not just registrations.
Open pipeline in the room, by stage illustrative
Opportunities created before the event date and still open on it · attending accounts vs registered no-show accounts
Accounts that attendedRegistered, no one came
Top accounts in the room
Ranked by attendees · ARR from the Dallas analysis; open pipeline illustrative until the query runs
| Account | Industry | ARR (est.) | Attended / reg. | Show rate | Open pipeline illustrative |
|---|---|---|---|---|---|
| Kiewit | Construction | $4.6M | 10 / 11 | 91% | $3.2M |
| Lina T. Ramey & Assoc. | Engineering | $348K | 10 / 14 | 71% | $0.4M |
| BNSF Railway | Rail & transit | $267K | 10 / 10 | 100% | $0.9M |
| Ferrovial | Construction | $1.0M | 9 / 20 | 45% | $1.6M |
| Black & Veatch | Engineering | $7.9M | 8 / 12 | 67% | $4.1M |
| AtkinsRéalis | Engineering | $11.0M | 8 / 15 | 53% | $5.3M |
| STV | Engineering | $3.9M | 7 / 11 | 64% | $1.2M |
| Jacobs | Engineering | $18.2M | 7 / 13 | 54% | $6.8M |
| AECOM | Engineering | $20.1M | 6 / 15 | 40% | $7.4M |
| Burns & McDonnell | Engineering | $7.4M | 6 / 8 | 75% | $2.2M |
How the event numbers are defined
Same rules for every AI Pillar event so results compare
Registered / attendedDistinct people on the event campaign (Salesforce members, Cvent/Splash upload). Attended = member status Attended / In person / Virtual. De-duplicated by Contact/Lead and email; test and duplicate records removed first.
External vs Bentley@bentley.com registrants are reported as their own cohort and excluded from external show rate, accounts and ARR.
Accounts & ARR in the roomAttendees rolled up to the ultimate-parent account; ARR = Salesforce ARR field on that account (already annualized). Existing customer value, not event revenue.
Open pipeline in the roomOpportunities at those accounts with Created Date before the event date and not closed by it (close date on/after the event, or still open). Stage and amount are today’s values until opportunity history is in Snowflake.
Pipeline created afterOpportunities at attending accounts created within 90 days after the event (same window as the progression rule).
Where it fitsEach event is a child campaign under an AI Pillar sub-theme, so the overview row and this page fill in from the same campaign tree — no event-specific build.
How the Data Layer Comes Together Proposed end-to-end integration architecture
01 · Source Systems
Raw observations and transactional records
SalesforceCampaigns, members, contact/lead IDs, opportunities
Marketo + Event PlatformsEmail, forms, webinars, Cvent/Splash/ON24
GA4 + Search/LLMContent actions, sessions, referrals, web/AI visibility
PR, Social, PIKC, eStoreMentions, engagement, internal enablement, sales
→
02 · Snowflake PRESENTATION
Verified objects the model reads today
MART · SalesforceDIM_CAMPAIGN · FCT_CAMPAIGN_MEMBERSHIP · DIM_CONTACT / DIM_LEAD · FCT_OPPORTUNITY
EDW · Marketo + ON24MARKETO_ACTIVITIES_EMAIL_* · MARKETO_EMAILS / LEADS · EVENT / REGISTRANT / ATTENDEE
MART_MARKETING · GA4DAILY_GA4_PAGES_REPORT (AI Hub + AI blog paths, AI referrers)
EDW · Paid mediaMEDIA_GOOGLE / BING / LINKEDIN / FB_AD_ANALYTICS · MEDIA_LOOKUPS
→
03 · Power BI Semantic Model
Native SQL per table · campaign tree · guardrails in DAX
Campaign Tree + Membersrecursive walk from the four sub-themes · Responded · New vs Existing
Email · Webinar · Web Dailyrecipient × email · registrant × webinar · day × page × source
Paid Daily + Paid CampaignsSnowflake spend joined to the live list (DirectQuery) on campaign ID
Pipelineprimary campaign in tree or account touch before creation
→
04 · Executive / Diagnostic Views
Measures in Power BI, consistent across pages
Reach & Visibilityunique known people; SOV; LLM; web
Engagement & Progressionqualified actions; known contacts; MQLs
Commercial Impactdistinct intent; influenced pipeline; revenue
Dashboard & drilldownsexecutive scorecard + channel / source pages
Sources verified in Snowflake (PRESENTATION) on 10/20/26 and wired into the live Power BI model. The full SQL, campaign hierarchy, data flow diagram and DAX are on the Build Spec & SQL page. The illustrative charts on this wireframe still use sample data.
Data Model: Required Keys & Fact Grain
Trace any metric or chart Pick or search a KPI, chart or table to see what it answers, how it’s calculated, the joins, the DAX and a query you can run in Snowflake
Data Quality Gates Before Production
Campaign hierarchyEvery event/digital campaign maps to one AI Pillar child and parent Campaign ID. Avoid name-based matching when stable IDs exist.
Person resolution & dedupeMap Marketo lead, event registrant and Salesforce Contact/Lead to one governed person key; retain anonymous activity separately.
Response-status definitionNormalize registered vs attended, responded vs non-responded, known content engagement and qualified MQL transitions.
Timing & attributionProgression must follow first engagement; opportunity influence requires an earlier qualifying touch within a governed lookback window.
Different measure grainsPeople ≠ content actions ≠ pageviews ≠ media impressions ≠ opportunities ≠ orders. Never combine units in one sum.
Finance & cancellationsCompleted eStore sales use canonical order ID, completed status, net amount and the finance-approved revenue recognition date.
Sample Data Inspector See normalized records behind the illustrations
First 10 rows of the current filtered sample; data generated deterministically in the browser
Build Spec & SQL How the live Power BI model connects Snowflake, the campaign hierarchy and every system — for the developer or anyone picking this up later
Metric & chart tracer Every KPI and chart on every page → source system → Snowflake objects → joins → Power BI table → DAX → check query
One source of truthEverything is read from Snowflake PRESENTATION with native SQL, except the paid campaign list, which is read live from Nischal’s Power BI model.
Hierarchy drives scopeSalesforce campaign parent IDs decide what counts as AI Pillar. Add a child campaign under a sub-theme and it flows in on the next refresh.
Parameters, not editsCampaign IDs, email IDs, webinar IDs and the paid alignment value are parameters, so the same model works for the next campaign.
Units never mixedPeople, actions, pageviews, impressions and dollars stay in separate measures. vs prior period compares the same-length window before.
Data flow · source to report
What feeds each Power BI table, and where it lands in Snowflake first
Campaign hierarchy · how everything rolls up to the AI Pillar
Salesforce campaigns are the spine; every other system attaches to a child campaign
How each system attaches to the hierarchy
The join key that ties each source to a child campaign, and the KPIs it feeds
| System | How it attaches | Key | Feeds | Status |
|---|---|---|---|---|
| Salesforce events (Cvent / Splash upload) | Child campaign under a sub-theme; attendees uploaded as members with status | Campaign Parent ID → sub-theme | Members, Reach, Engaged, New/Existing, MQLs, Pipeline | Needs set-up |
| Marketo email programs | Mailing IDs listed in {{EmailIds}} today; target: program synced to an SFDC child campaign | Mailing ID (today) → SFDC Campaign ID (target) | Email, Reach, Email CTO/CTR, AAG reach | Ready |
| Marketo gated content / forms | Form campaign as SFDC child; Responded flag set on fill | SFDC Campaign ID | Content form fills, TLE, Engaged, Progression | Needs set-up |
| ON24 webinars | Event IDs in {{WebinarEventIds}}; target: ON24 event linked to an SFDC child campaign | Event ID → SFDC Campaign ID | Webinar registrations / attendance, TLE | Ready · none scheduled |
| GA4 web | Page-path rules for AI Hub / AI blog; utm_campaign carries the child campaign key on campaign links | Path rules · utm_campaign | LLM & web visibility, AI referrals, top content, TLE | Ready |
| Paid media | Row in the paid campaign Microsoft List: Campaign Alignment = “Pillar - AI” + SFDC Child Campaign ID | Platform campaign ID → list → SFDC child | Paid impressions / clicks / conversions (diagnostic) | Ready |
| Opportunities | Primary campaign in the tree, or account member touch before creation; later Salesforce Campaign Influence | Primary campaign key · ultimate account key | Influenced pipeline & opportunities | Needs definition |
| PR / SOV, analyst briefings, social, PIKC, ClickUp (events, UG, CAB) | No feed yet — placeholder measures show “—” | — | Share of voice, PR impressions, social, video, PIKC, activity counts | Manual entry |
Connection & parameters
Power Query parameters used by every query (shown as {{Name}} in the SQL)
| Parameter | Current value | Purpose |
|---|---|---|
SnowflakeServer | vkelqpo-bird_prod.snowflakecomputing.com | Snowflake account |
SnowflakeWarehouse | REPORTING_WH__XS | Warehouse used by the refresh |
StartDate | 2024-07-01 | Earliest date loaded (gives prior-period history) |
PillarRootId | 701Wj00001uusRZIAY | AI Pillar root campaign (reference) |
SubThemeIds | '701Wj00001x4BSLIA2', '701Wj00001x5JE3IAM', '701Wj00001x4pG1IAI', '701Wj00001x5VofIAE' | The four sub-themes the tree starts from |
EmailIds | 31004,31138 | Marketo mailing IDs in scope |
WebinarEventIds | 0 | ON24 event IDs in scope (none scheduled yet) |
PaidAlignment | Pillar - AI | Campaign Alignment value in the paid campaign list |
Model, refresh & access
How the published model behaves
| Storage | Import for every Snowflake table; DirectQuery for Paid Campaigns (composite model) |
| Relationships | All facts → Date[Date] (single direction) · Members → Campaign Tree on campaign ID · Paid Daily → Paid Campaigns on platform campaign ID |
| Calculated tables | Date (1 Jul 2024 → today), Metric Catalog, Funnel Stage, Team Headline, TLE Type |
| Report filter | Campaign window: Date ≥ 1 Oct 2025 (earlier data kept for prior-period comparisons) |
| Refresh | Daily ~7 AM CT on the published dataset in My workspace |
| Credentials | Snowflake: OAuth2 (Maurice) · Paid Media updated: OAuth2 — both signed in on the dataset settings page |
| Native queries | Desktop asks to approve each new SQL text once; the service runs them without prompting |
| Project files | Power BI project (.pbip): report in PBIR folders, model in TMDL — diffable and versionable |
Power BI tables · grain, source and logic
One line per model table; full SQL below
| Table | Grain | Snowflake objects | Logic | Status |
|---|---|---|---|---|
| Campaign Tree | One row per campaign in the AI Pillar tree (sub-themes + every child below them) | MART.DIM_CAMPAIGN | Recursive CTE from the four sub-theme IDs down the Campaign Parent ID chain (max 6 levels). Tags each node with its sub-theme, level and Sub-theme / Tactic. | Ready · 0 rows today |
| Members | One row per campaign member on any campaign in the tree | MART.FCT_CAMPAIGN_MEMBERSHIP · DIM_CAMPAIGN_MEMBER_STATUS · DIM_CONTACT · DIM_LEAD · FCT_LEAD | Same tree CTE, joined to membership. Responded = 1 when Has Campaign Member Responded is TRUE/YES/1. New vs Existing = person created between 1 day before and 14 days after the member record. Region from contact or lead. | Ready · 0 rows today |
| Pipeline | One row per opportunity influenced by the pillar | MART.FCT_OPPORTUNITY · DIM_OPPORTUNITY · FCT_CAMPAIGN_MEMBERSHIP · DIM_CAMPAIGN | Influenced when (a) the opportunity’s primary campaign is in the tree, or (b) any member of the same ultimate account joined a tree campaign on or before the opportunity create date. Records which rule fired and the touch campaign. | Ready · 0 rows today |
| One row per recipient × email | EDW.MARKETO_ACTIVITIES_EMAIL_SEND / DELIVERED / OPEN / CLICK · MARKETO_EMAILS · MARKETO_LEADS | Mailings filtered to {{EmailIds}}. Delivered / opened / clicked flags per recipient (bot clicks excluded). Segment = Internal enablement when the email address is @bentley.com, otherwise External market. | Ready | |
| Webinar | One row per registrant × webinar | EDW.EVENT · REGISTRANT · ATTENDEE (ON24) | Events filtered to {{WebinarEventIds}} (current rows only). Attended when the registrant email appears in ATTENDEE; minutes = live + archive. | Ready · none scheduled |
| Web Daily | One row per day × page × source × campaign | MART_MARKETING.DAILY_GA4_PAGES_REPORT | AI pages = path contains infrastructure-ai, /category/ai/, or a /news/ or /blog/ path with -ai- / artificial-intelligence. Page group = AI Hub vs AI Blog & News. AI referral = session source ChatGPT/OpenAI, Gemini, Perplexity, Copilot or Claude. | Ready |
| Paid Daily | One row per day × platform campaign | EDW.MEDIA_GOOGLE / BING / LINKEDIN / FB_AD_ANALYTICS · MEDIA_LOOKUPS | All four platforms unioned from {{StartDate}}. The AI Pillar filter is applied in the measures through the live Paid Campaigns list (Campaign Alignment = “Pillar - AI”), joined on platform campaign ID. | Ready |
| Paid Campaigns | One row per paid platform campaign (DirectQuery) | Power BI semantic model “Paid Media updated” (Marketing Analytics Dashboards) · table Paid Media Campaign Metadata | Read live with the viewer’s sign-in — no copy, no XMLA needed. Columns: CAMPAIGNID, Platform, Campaign Name, Campaign Alignment, SFDC Child Campaign ID, Funnel Stage, Region. | Live (DirectQuery) |
SQL behind each table Exactly what the model runs · parameters in pink · click to expand, copy to run in Snowflake (replace the parameters)
Campaign Tree Ready · 0 rows today
- Grain
- One row per campaign in the AI Pillar tree (sub-themes + every child below them)
- Reads
- MART.DIM_CAMPAIGN
- Logic
- Recursive CTE from the four sub-theme IDs down the Campaign Parent ID chain (max 6 levels). Tags each node with its sub-theme, level and Sub-theme / Tactic.
with recursive
tree (campaign_id, sub_theme_id, lvl) as (
select "Campaign ID",
"Campaign ID",
1
from PRESENTATION.MART.DIM_CAMPAIGN
where "Campaign ID" in ({{SubThemeIds}})
union all
select c."Campaign ID",
t.sub_theme_id,
t.lvl + 1
from PRESENTATION.MART.DIM_CAMPAIGN c
join tree t on c."Campaign Parent ID" = t.campaign_id
where t.lvl < 6
)
select t.campaign_id,
c."Campaign Name" as campaign_name,
c."Campaign Parent ID" as parent_campaign_id,
t.sub_theme_id,
s."Campaign Name" as sub_theme,
t.lvl as tree_level,
iff(t.lvl = 1, 'Sub-theme', 'Tactic') as node_type,
c."Campaign Type" as campaign_type,
c."Campaign Status" as campaign_status,
c."Campaign Start Date" as start_date,
c."Campaign End Date" as end_date,
'{{PillarRootId}}' as pillar_root_id
from tree t
join PRESENTATION.MART.DIM_CAMPAIGN c on c."Campaign ID" = t.campaign_id
join PRESENTATION.MART.DIM_CAMPAIGN s on s."Campaign ID" = t.sub_theme_id
Members Ready · 0 rows today
- Grain
- One row per campaign member on any campaign in the tree
- Reads
- MART.FCT_CAMPAIGN_MEMBERSHIP · DIM_CAMPAIGN_MEMBER_STATUS · DIM_CONTACT · DIM_LEAD · FCT_LEAD
- Logic
- Same tree CTE, joined to membership. Responded = 1 when Has Campaign Member Responded is TRUE/YES/1. New vs Existing = person created between 1 day before and 14 days after the member record. Region from contact or lead.
with recursive
tree (campaign_id, sub_theme_id, lvl) as (
select "Campaign ID",
"Campaign ID",
1
from PRESENTATION.MART.DIM_CAMPAIGN
where "Campaign ID" in ({{SubThemeIds}})
union all
select c."Campaign ID",
t.sub_theme_id,
t.lvl + 1
from PRESENTATION.MART.DIM_CAMPAIGN c
join tree t on c."Campaign Parent ID" = t.campaign_id
where t.lvl < 6
)
select m.PK_CAMPAIGN_MEMBERSHIP_KEY as membership_key,
m."Campaign ID" as campaign_id,
t.sub_theme_id,
m.FK_CONTACT_OR_LEAD_KEY as person_key,
iff(m.FK_CONTACT_KEY is not null, 'Contact', 'Lead') as person_type,
m.FK_ULTIMATE_KEY as account_key,
st."Campaign Member Status" as member_status,
iff(upper(m."Has Campaign Member Responded") in ('TRUE','YES','1'), 1, 0) as responded,
m."Campaign Member Created At"::date as member_created_date,
m."Campaign Member First Responded At"::date as first_responded_date,
coalesce(to_date(to_char(ct.FK_CONTACT_CREATED_DATE_KEY), 'YYYYMMDD'), to_date(to_char(fl.FK_LEAD_CREATED_DATE_KEY), 'YYYYMMDD')) as person_created_date,
iff(datediff('day', coalesce(to_date(to_char(ct.FK_CONTACT_CREATED_DATE_KEY), 'YYYYMMDD'), to_date(to_char(fl.FK_LEAD_CREATED_DATE_KEY), 'YYYYMMDD')), m."Campaign Member Created At"::date) between -1 and 14, 'New', 'Existing') as new_vs_existing,
coalesce(ct."Contact Organizational Region", ld."Lead Organizational Region") as region
from PRESENTATION.MART.FCT_CAMPAIGN_MEMBERSHIP m
join tree t on t.campaign_id = m."Campaign ID"
left join PRESENTATION.MART.DIM_CAMPAIGN_MEMBER_STATUS st on st.SK_CAMPAIGN_MEMBER_STATUS_KEY = m.FK_CAMPAIGN_MEMBER_STATUS_KEY
left join PRESENTATION.MART.DIM_CONTACT ct on ct.SK_CONTACT_KEY = m.FK_CONTACT_KEY
left join PRESENTATION.MART.DIM_LEAD ld on ld.SK_LEAD_KEY = m.FK_LEAD_KEY
left join PRESENTATION.MART.FCT_LEAD fl on fl.FK_LEAD_KEY = m.FK_LEAD_KEY
Pipeline Ready · 0 rows today
- Grain
- One row per opportunity influenced by the pillar
- Reads
- MART.FCT_OPPORTUNITY · DIM_OPPORTUNITY · FCT_CAMPAIGN_MEMBERSHIP · DIM_CAMPAIGN
- Logic
- Influenced when (a) the opportunity’s primary campaign is in the tree, or (b) any member of the same ultimate account joined a tree campaign on or before the opportunity create date. Records which rule fired and the touch campaign.
with recursive
tree (campaign_id, sub_theme_id, lvl) as (
select "Campaign ID",
"Campaign ID",
1
from PRESENTATION.MART.DIM_CAMPAIGN
where "Campaign ID" in ({{SubThemeIds}})
union all
select c."Campaign ID",
t.sub_theme_id,
t.lvl + 1
from PRESENTATION.MART.DIM_CAMPAIGN c
join tree t on c."Campaign Parent ID" = t.campaign_id
where t.lvl < 6
),
tree_keys as (
select dc.SK_CAMPAIGN_KEY,
t.campaign_id,
t.sub_theme_id
from tree t
join PRESENTATION.MART.DIM_CAMPAIGN dc on dc."Campaign ID" = t.campaign_id
),
acct_touch as (
select m.FK_ULTIMATE_KEY,
min(m."Campaign Member Created At"::date) as first_touch_date,
min_by(t.campaign_id, m."Campaign Member Created At") as first_touch_campaign
from PRESENTATION.MART.FCT_CAMPAIGN_MEMBERSHIP m
join tree t on t.campaign_id = m."Campaign ID"
where m.FK_ULTIMATE_KEY is not null
group by 1
)
select o.PK_OPPORTUNITY_KEY as opportunity_key,
d."Opportunity ID" as opportunity_id,
d."Opportunity Name" as opportunity_name,
d."Opportunity Stage" as stage,
d."Opportunity Status" as status,
o."Opportunity Create Date" as created_date,
o."Opportunity Close Date" as close_date,
o."Opportunity Amount USD" as amount_usd,
o.FK_ULTIMATE_KEY as account_key,
iff(pk.campaign_id is not null, 'Primary campaign', 'Account touch before creation') as touch_rule,
coalesce(pk.campaign_id, at.first_touch_campaign) as touch_campaign_id,
at.first_touch_date
from PRESENTATION.MART.FCT_OPPORTUNITY o
join PRESENTATION.MART.DIM_OPPORTUNITY d on d.SK_OPPORTUNITY_KEY = o.FK_OPPORTUNITY_KEY
left join tree_keys pk on pk.SK_CAMPAIGN_KEY = o.FK_PRIMARY_CAMPAIGN_KEY
left join acct_touch at on at.FK_ULTIMATE_KEY = o.FK_ULTIMATE_KEY and at.first_touch_date <= o."Opportunity Create Date"
where o."Opportunity Create Date" >= '{{StartDate}}' and (pk.campaign_id is not null or at.FK_ULTIMATE_KEY is not null)
Email Ready
- Grain
- One row per recipient × email
- Reads
- EDW.MARKETO_ACTIVITIES_EMAIL_SEND / DELIVERED / OPEN / CLICK · MARKETO_EMAILS · MARKETO_LEADS
- Logic
- Mailings filtered to {{EmailIds}}. Delivered / opened / clicked flags per recipient (bot clicks excluded). Segment = Internal enablement when the email address is @bentley.com, otherwise External market.
with
s as (
select LEADID,
MAILINGID,
min(ACTIVITY_DATE)::date as sent_date
from PRESENTATION.EDW.MARKETO_ACTIVITIES_EMAIL_SEND
where MAILINGID in ({{EmailIds}})
group by 1,2
),
d as (
select distinct LEADID,
MAILINGID
from PRESENTATION.EDW.MARKETO_ACTIVITIES_EMAIL_DELIVERED
where MAILINGID in ({{EmailIds}})
),
o as (
select distinct LEADID,
MAILINGID
from PRESENTATION.EDW.MARKETO_ACTIVITIES_EMAIL_OPEN
where MAILINGID in ({{EmailIds}})
),
c as (
select distinct LEADID,
MAILINGID
from PRESENTATION.EDW.MARKETO_ACTIVITIES_EMAIL_CLICK
where MAILINGID in ({{EmailIds}}) and not coalesce(IS_BOT_ACTIVITY, false)
)
select s.LEADID as marketo_lead_id,
lower(ml.EMAIL) as email,
s.MAILINGID as email_id,
e.NAME as email_name,
s.sent_date,
iff(d.LEADID is null, 0, 1) as delivered,
iff(o.LEADID is null, 0, 1) as opened,
iff(c.LEADID is null, 0, 1) as clicked,
iff(lower(ml.EMAIL) like '%@bentley.com', 'Internal enablement', 'External market') as segment
from s
left join d on d.LEADID = s.LEADID and d.MAILINGID = s.MAILINGID
left join o on o.LEADID = s.LEADID and o.MAILINGID = s.MAILINGID
left join c on c.LEADID = s.LEADID and c.MAILINGID = s.MAILINGID
left join PRESENTATION.EDW.MARKETO_EMAILS e on e.ID = s.MAILINGID
left join PRESENTATION.EDW.MARKETO_LEADS ml on ml.ID = s.LEADID
Webinar Ready · none scheduled
- Grain
- One row per registrant × webinar
- Reads
- EDW.EVENT · REGISTRANT · ATTENDEE (ON24)
- Logic
- Events filtered to {{WebinarEventIds}} (current rows only). Attended when the registrant email appears in ATTENDEE; minutes = live + archive.
select e.EVENT_ID as event_id,
e.EVENT_NAME as event_name,
e.EVENT_LIVE_START::date as live_date,
lower(r.EMAIL) as email,
r.CREATE_TIMESTAMP::date as registered_date,
r.UTM_CAMPAIGN as utm_campaign,
iff(a.EMAIL is null, 0, 1) as attended,
a.ENGAGEMENT_SCORE as engagement_score,
coalesce(a.LIVE_MINUTE, 0) + coalesce(a.ARCHIVE_MINUTE, 0) as minutes_viewed
from PRESENTATION.EDW.EVENT e
join PRESENTATION.EDW.REGISTRANT r on r.EVENT_ID = e.EVENT_ID and r.IS_CURRENT_ROW
left join PRESENTATION.EDW.ATTENDEE a on a.EVENT_ID = e.EVENT_ID and lower(a.EMAIL) = lower(r.EMAIL) and a.IS_CURRENT_ROW
where e.IS_CURRENT_ROW and e.EVENT_ID in ({{WebinarEventIds}})
Web Daily Ready
- Grain
- One row per day × page × source × campaign
- Reads
- MART_MARKETING.DAILY_GA4_PAGES_REPORT
- Logic
- AI pages = path contains infrastructure-ai, /category/ai/, or a /news/ or /blog/ path with -ai- / artificial-intelligence. Page group = AI Hub vs AI Blog & News. AI referral = session source ChatGPT/OpenAI, Gemini, Perplexity, Copilot or Claude.
select "Event Date" as event_date,
case when ("Path" ilike '%/news/%' or "Path" ilike '%/blog/%') then 'AI Blog & News' when "Path" ilike '%infrastructure-ai%' then 'AI Hub' else 'AI Blog & News' end as page_group,
"Path" as path,
"Country" as country,
"Session Default Channel Group" as channel_group,
"Session Source" as session_source,
"Session Campaign Name" as utm_campaign,
iff("Session Source" ilike any ('%chatgpt%','%openai%','%perplexity%','%copilot%','%gemini%','claude%'), 1, 0) as is_ai_referral,
case when "Session Source" ilike any ('%chatgpt%','%openai%') then 'ChatGPT' when "Session Source" ilike '%gemini%' then 'Gemini' when "Session Source" ilike '%perplexity%' then 'Perplexity' when "Session Source" ilike '%copilot%' then 'Copilot' when "Session Source" ilike 'claude%' then 'Claude' end as ai_assistant,
sum("Pageviews") as pageviews,
sum("Users") as users,
sum("Sessions") as sessions,
sum("Engaged Sessions") as engaged_sessions
from PRESENTATION.MART_MARKETING.DAILY_GA4_PAGES_REPORT
where "Event Date" >= '{{StartDate}}' and ( "Path" ilike '%infrastructure-ai%' or "Path" ilike '%/category/ai/%' or (("Path" ilike '%/news/%' or "Path" ilike '%/blog/%') and ("Path" ilike '%-ai-%' or "Path" ilike '%-ai/%' or "Path" ilike '%artificial-intelligence%')) )
group by 1,2,3,4,5,6,7,8,9
Paid Daily Ready
- Grain
- One row per day × platform campaign
- Reads
- EDW.MEDIA_GOOGLE / BING / LINKEDIN / FB_AD_ANALYTICS · MEDIA_LOOKUPS
- Logic
- All four platforms unioned from {{StartDate}}. The AI Pillar filter is applied in the measures through the live Paid Campaigns list (Campaign Alignment = “Pillar - AI”), joined on platform campaign ID.
with
paid as (
select 'Google' as platform,
REPORTED_DATE,
CAMPAIGNID,
IMPRESSIONS,
CLICKS,
COST,
CONVERSIONS,
null::number as leads,
VIDEO_VIEWS
from PRESENTATION.EDW.MEDIA_GOOGLE_AD_ANALYTICS
union all
select 'Microsoft Advertising',
REPORTED_DATE,
CAMPAIGNID,
IMPRESSIONS,
CLICKS,
COST,
CONVERSIONS,
null,
null
from PRESENTATION.EDW.MEDIA_BING_AD_ANALYTICS
union all
select 'LinkedIn',
REPORTED_DATE,
CAMPAIGNID,
IMPRESSIONS,
CLICKS,
COST,
CONVERSIONS,
LEADS,
VIDEO_VIEWS
from PRESENTATION.EDW.MEDIA_LINKEDIN_AD_ANALYTICS
union all
select 'Meta',
REPORTED_DATE,
CAMPAIGNID,
IMPRESSIONS,
CLICKS,
COST,
null,
LEADS,
VIDEO_VIEWS
from PRESENTATION.EDW.MEDIA_FB_AD_ANALYTICS
)
select p.platform,
p.REPORTED_DATE as report_date,
p.CAMPAIGNID as platform_campaign_id,
max(l.CAMPAIGN_NAME) as platform_campaign_name,
sum(p.IMPRESSIONS) as impressions,
sum(p.CLICKS) as clicks,
sum(p.COST) as cost,
sum(p.CONVERSIONS) as conversions,
sum(p.leads) as leads,
sum(p.VIDEO_VIEWS) as video_views
from paid p
left join (select distinct CAMPAIGNID, CAMPAIGN_NAME from PRESENTATION.EDW.MEDIA_LOOKUPS) l on l.CAMPAIGNID = p.CAMPAIGNID
where p.REPORTED_DATE >= '{{StartDate}}'
group by 1,2,3
Paid Campaigns Live (DirectQuery)
- Source
- Power BI semantic model Paid Media updated in Marketing Analytics Dashboards (owner: Nischal)
- Mode
- DirectQuery (composite model) — the AI Pillar model holds no copy
- Upstream list
- SharePoint Microsoft List Paid Media Campaign Metadata (site 2020MarketingPlans) — open list. Media owners keep Campaign Alignment and SFDC Child Campaign ID filled in there; that ID is what places a paid campaign in the AI Pillar campaign tree.
- Why not SQL
- The list is maintained in a Microsoft List behind Nischal’s model; that workspace is not on Premium capacity, so an import refresh through XMLA fails in the service. DirectQuery works on Pro.
let
Source = AnalysisServices.Database(
"powerbi://api.powerbi.com/v1.0/myorg/Marketing%20Analytics%20Dashboards",
"Paid Media updated"),
Cubes = Table.Combine(Source[Data]),
Cube = Cubes{[Id="Model", Kind="Cube"]}[Data]
in
Cube
// table: Paid Media Campaign Metadata → local table Paid Campaigns
Event impact · open pipeline in the room Needs set-up
- Grain
- Event × account turnout × opportunity stage
- Reads
- MART.DIM_CAMPAIGN · FCT_CAMPAIGN_MEMBERSHIP · DIM_CAMPAIGN_MEMBER_STATUS · FCT_OPPORTUNITY · DIM_OPPORTUNITY
- Logic
- Accounts = ultimate parents of the event’s members (attended if any member attended). Open in the room = created before the event date and not closed by it. Stage/amount are current values until opportunity history lands in Snowflake.
-- Open pipeline in the room for one event campaign
-- (replace {{EventCampaignId}}; in the model this runs for every event campaign in the AI Pillar tree)
with ev as (
select "Campaign ID" as campaign_id,
"Campaign Start Date"::date as event_date
from PRESENTATION.MART.DIM_CAMPAIGN
where "Campaign ID" = '{{EventCampaignId}}'
),
accts as (
select m.FK_ULTIMATE_KEY as account_key,
max(iff(st."Campaign Member Status" ilike 'attended%', 1, 0)) as attended
from PRESENTATION.MART.FCT_CAMPAIGN_MEMBERSHIP m
join ev on ev.campaign_id = m."Campaign ID"
left join PRESENTATION.MART.DIM_CAMPAIGN_MEMBER_STATUS st on st.SK_CAMPAIGN_MEMBER_STATUS_KEY = m.FK_CAMPAIGN_MEMBER_STATUS_KEY
where m.FK_ULTIMATE_KEY is not null
group by 1
)
select iff(a.attended = 1, 'Attended', 'Registered, no one came') as account_turnout,
d."Opportunity Stage" as stage,
count(distinct o.PK_OPPORTUNITY_KEY) as open_opps,
sum(o."Opportunity Amount USD") as open_pipeline_usd
from PRESENTATION.MART.FCT_OPPORTUNITY o
join PRESENTATION.MART.DIM_OPPORTUNITY d on d.SK_OPPORTUNITY_KEY = o.FK_OPPORTUNITY_KEY
join accts a on a.account_key = o.FK_ULTIMATE_KEY
cross join ev
where o."Opportunity Create Date" < ev.event_date
and (o."Opportunity Close Date" >= ev.event_date or d."Opportunity Status" = 'Open')
group by 1, 2
order by 1, 4 desc;
-- Pipeline created within 90 days after the event, at attending accounts
-- ... same ev / accts CTEs ...
-- where a.attended = 1
-- and o."Opportunity Create Date" between ev.event_date and dateadd('day', 90, ev.event_date)
Measure logic (DAX) How the headline numbers are calculated in the model
Total audience reached (external)
[Email Recipients (External)] + [Event & Community People]
-- Email Recipients (External)
CALCULATE(DISTINCTCOUNT('Email'[EMAIL]), 'Email'[DELIVERED] = 1, 'Email'[SEGMENT] = "External market")
-- Event & Community People
COALESCE(DISTINCTCOUNT('Members'[PERSON_KEY]), 0)
LLM & web visibility
[AI Hub Pageviews] + [AI Blog Pageviews]
-- AI Hub Pageviews
CALCULATE(SUM('Web Daily'[PAGEVIEWS]), 'Web Daily'[PAGE_GROUP] = "AI Hub")
-- AI Referral Sessions (shown separately, never summed with pageviews)
CALCULATE(SUM('Web Daily'[SESSIONS]), 'Web Daily'[IS_AI_REFERRAL] = 1)
Thought leadership engagement
[Webinar Registrations]
+ CALCULATE(COUNTROWS('Members'), 'Members'[RESPONDED] = 1)
+ CALCULATE(SUM('Web Daily'[ENGAGED_SESSIONS]), 'Web Daily'[PAGE_GROUP] = "AI Hub")
-- video views, PIKC downloads, social and form fills join as their sources land
Engaged known contacts
COALESCE(CALCULATE(DISTINCTCOUNT('Members'[PERSON_KEY]), 'Members'[RESPONDED] = 1), 0)
Influenced pipeline
COALESCE(SUM('Pipeline'[AMOUNT_USD]), 0)
Paid media (diagnostic)
CALCULATE(SUM('Paid Daily'[IMPRESSIONS]),
KEEPFILTERS('Paid Campaigns'[CAMPAIGN_ALIGNMENT] = "Pillar - AI"))
vs prior period (every KPI)
VAR _min = MIN('Date'[Date])
VAR _max = MAX('Date'[Date])
VAR _d = INT(_max - _min) + 1
RETURN CALCULATE([<KPI>], DATESBETWEEN('Date'[Date], _min - _d, _min - 1))
-- label: ▲/▼ % change (points for rates); green up, red down; "no prior period" when blank
Open items before the numbers are final
Child campaigns under the sub-themesThe four sub-themes have no children and no members in Snowflake today, so events, engaged contacts, progression, MQLs and pipeline read 0. Parent every AI event, webinar, form and paid child campaign to one sub-theme in Salesforce.
De-duplicate people across channelsAudience = email recipients + campaign members. Email is keyed on lower-case email, members on Contact/Lead key, so one person in both is counted twice today. Fix: resolve Marketo lead → SFDC Contact/Lead (MARKETO_LEADS SFDC ID, else email) and count distinct person keys; drop test records.
Responded flag on event statusesEngaged known contacts relies on Has Campaign Member Responded. Set it for Attended / Visited booth / Meeting held statuses on event campaigns.
MQL date in SnowflakeMQLs from the campaign need the MQL timestamp on lead/contact; not in MART today.
Intent requestsDemo / trial / contact-sales / EA-LA need the product CTA form list mapped to request type (trial and EA/LA source unknown).
Credentials for refreshSnowflake (OAuth2) and Paid Media updated (OAuth2) must be signed in on the published dataset before the daily refresh and the Team & Funnel page work in the service.
Reuse this for the next campaign
Build the campaign tree in SalesforceCreate the root and sub-theme campaigns; parent every tactic campaign (events, webinars, forms, email, paid) to a sub-theme.
Set member statuses + RespondedUse the standard status set and tick Responded for the qualifying statuses so engagement and progression work.
Tag the other systemsMarketo programs synced to SFDC children; ON24 events linked; paid list rows with the alignment value + SFDC Child Campaign ID; utm_campaign = child campaign key.
Change the parametersCopy the .pbip, then set SubThemeIds, PillarRootId, EmailIds, WebinarEventIds, PaidAlignment and StartDate. No SQL edits needed.
Refresh + sign inRefresh in Desktop (approve the native queries once), publish, sign in Snowflake + Paid Media credentials, schedule the daily refresh.
ValidateCheck tree row count, member and email counts against Salesforce / Marketo, and de-dup people before sharing numbers.
Executive KPI Definitions
Individual Metrics → Executive Outcomes Directly mapped from the workbook’s KPI Mapping Detail tab
“Proposed / needs validation” means a business rule or integration has not been verified. The demo numbers reflect illustrative model behavior only; they do not imply data readiness.
Feedback summary Every comment left on this wireframe, by page
Reference design: Bentley Sale Week dashboard screenshot · KPI source: AI Pillar Campaign Metrics workbook (Sheet1, Executive Scorecard, KPI Mapping Detail) · All displayed numbers, products, opportunities, and accounts are synthetic. Filters affect simulated data and metric measures; nothing is queried from production systems.