Impact Report / Account Intelligence & Pipeline
Pipeline is visible through DemandSense now. HubSpot is connected live through the DemandSense sync, so every ad-touched company and CRM lead surfaces in the three tier attribution model in section 02b. The one field the client has not populated is the deal stage: it is blank on 417 of 418 records and the close date on all 418. So we can see which companies the ads influenced, but not yet whether any of them closed. Every figure here is lead level.
The account list includes parties that cannot buy. Northwind Commerce itself is the single highest scoring LinkedIn engaged account. Impactable appears as a dual signal account. Competitor 27, a named direct competitor from the April white space analysis, sits at very high engagement. Amazon sits at very high engagement. Nine distinct defects are documented in section 03 and every one inflates the ABM numbers upward.
Consequence: no account list from this report should be handed to a sales motion until the exclusion pass in section 06 is applied. Two named people in the identified set carry LinkedIn URLs belonging to different people.
Reached through to customer. The ladder is honest about where it stops: identification and CRM match are both visible through DemandSense, and deal outcomes appear the moment HubSpot deal stages are populated.
n/a means the HubSpot deal stage is not yet populated. It is not a measured zero and must never be presented as one. 830 account linked people against 1,815 individually identified people is itself a reconciliation gap between two DemandSense models, flagged in section 03.
Account progression rather than lead progression. Thresholds as specified: exposed at one impression, aware at three, engaged at three clicks, then MQL and opportunity.
All 3,928 companies with any delivery have 3 or more impressions. The ABM export only includes companies that received meaningful delivery, so the one-versus-three impression split cannot separate them. To make Aware meaningful, either raise the threshold materially, around 50 impressions puts roughly 390 accounts in tier, or pull the full delivery log including one and two impression companies. As built, treat Exposed and Aware as a single stage.
3,928 aware accounts become 289 engaged, 7.4%. Of those 289, 192 are ICP fit and 120 are on the ABM list, so the engaged tier is well qualified. Then 289 becomes 128 CRM leads. The bottleneck is engagement depth, not reach and not qualification, which is consistent with a middle funnel that has no closing offer behind it.
This is a single snapshot. Showing how awareness built across three months requires three Company Hub exports, and only one exists. Either the LinkedIn specialist pulls monthly snapshots, or it is approximated from the monthly company intelligence model and labelled derived. A derived progression presented as measured is exactly the claim that will not survive a question.
This is the most important table in the report. Engagement tier behaves as a rate, not a volume, so large enterprises land in the lower tiers despite consuming the most delivery.
| Tier▿ | Accounts▿ | % of accounts▿ | Impressions▿ | % of impressions▿ | Impr. per account▿ | Clicks▿ | Leads▿ | Click to lead▿ | Est. spend▿ |
|---|---|---|---|---|---|---|---|---|---|
| Total | 9,787 | 100% | 218,269 | 100% | 22 | 4,324 | 17 | 0.39% | $24,411 |
Estimated spend apportions the verified $24,411 by each tier's share of company resolved impressions. It is an estimate, not a measured per account cost, because LinkedIn does not expose spend at account grain. Treat the ranking as sound and the absolute dollars as indicative.
HubSpot is connected through DemandSense. Deal-stage fields are not yet populated, so this is lead-level signal pairing rather than pipeline value. Leads window 3 May to 30 July, ad activity window 3 March to 27 July.
| Tier | Definition | Leads | Basis |
|---|---|---|---|
| Tier 1 | CRM record carries the LinkedIn Ads label | 17 | Click verified. Set by the client's own tagging. |
| Tier 2 | Lead email domain matches an ad exposed company that clicked | 28 | Impression and click verified at company level. |
| Tier 3 | Lead domain matches an ad exposed company, impressions only | 82 | Ecosystem influence. Weakest tier, do not headline. |
| Tier 4 | No ad exposure found | 280 | Upper bound. The join sees only 55% of spend. |
| Total distinct leads | 407 | 11 duplicate records and 1 CRM test record removed from 418. |
| Account | Leads | Clicks | Impr | Est. cost | ABM | Website | CRM label |
|---|---|---|---|---|---|---|---|
| ICP Brand 66 | 1 | 0 | 59 | n/a | Yes | Yes | Yes |
| Enterprise Account 53 | 1 | 176 | 4,454 | $686 | n/a | Yes | n/a |
| ICP Brand 91 / ICP Brand 84 | 2 | 17 | 563 | $65 | Yes | Yes | n/a |
| ICP Brand 44 | 1 | 0 | 93 | n/a | Yes | Yes | n/a |
| ICP Brand 62 | 1 | 18 | 924 | $95 | Yes | n/a | n/a |
| ICP Brand 103 | 2 | 15 | 308 | $47 | n/a | n/a | n/a |
| ICP Brand 63 | 1 | 8 | 254 | $27 | n/a | n/a | n/a |
| ICP Brand 71 | 4 | 6 | 579 | $50 | n/a | n/a | n/a |
| ICP Brand 6 | 4 | 6 | 151 | $14 | n/a | n/a | n/a |
| ICP Brand 106 | 4 | 4 | 141 | $23 | Yes | n/a | n/a |
| ICP Brand 15 | 6 | 0 | 161 | n/a | Yes | n/a | n/a |
| ICP Brand 107 | 5 | 0 | 275 | n/a | n/a | n/a | n/a |
| ICP Brand 29 | 4 | 0 | 194 | n/a | n/a | n/a | n/a |
| ICP Brand 41 | 4 | 0 | 121 | n/a | Yes | n/a | n/a |
| ICP Brand 90 | 5 | 0 | 112 | n/a | Yes | n/a | n/a |
| ICP Brand 56 | 2 | 0 | 308 | n/a | n/a | n/a | n/a |
| ICP Brand 36 | 1 | 0 | 82 | n/a | Yes | n/a | n/a |
| ICP Brand 46 | 1 | 0 | 79 | n/a | Yes | n/a | n/a |
On the ABM list, exposed to ads, visited the website, and present in the CRM carrying the LinkedIn Ads label. Nothing else in 407 leads has all four. It is the single clearest example of the full funnel working end to end.
Enterprise Account 53, ICP Brand 103, ICP Brand 62, ICP Brand 63 and ICP Brand 73 are all 5,000+ employee companies and all appear as ad exposed CRM leads. Reports 1 and 2 both concluded enterprise does not convert. Without deal stages populated we cannot tell whether these are real opportunities or noise, so the enterprise reduction case is now weaker than stated and should be softened until outcomes arrive.
The DemandSense ABM signal set resolves $13,414 of $24,411 spend, 3,416 of 6,421 clicks and 157,210 of 230,251 impressions. Tier 4 is therefore an upper bound. A non match is not proof that a company was never served an ad.
Estimated cost is the company level spend recorded in the ABM export, not an allocation of the lead. Accounts shown are ICP fit and ad touched, deduplicated, and named by email domain because the company name field is unreliable on wrapped rows.
Nine defects found while assembling this report. All of them bias the account intelligence upward, which means every ABM figure is biased upward until these are excluded.
| # | Defect | Evidence | Consequence |
|---|---|---|---|
| 1 | Client self engagement | Northwind Commerce scores 20,706 with 969 organic engagements, ranking first overall. | Employees top their own leaderboard. Exclude. |
| 2 | Duplicate entity, same name | A second company resolves to Northwind Commerce on example.com, 10 sessions, with a named CTO. | Name matching merges two companies. |
| 3 | Agency contamination | Impactable scores 443 with 28 organic engagements, flagged dual signal. | Agency engagement counted as client demand. |
| 4 | Direct competitors engaged | Competitor 27 2,680 at very high tier. Competitor 1 117 impressions, 38 organic engagements, very high. Plus Competitor 9 and Competitor 10 with named executives. | Competitors inflate the top tier that carries the ABM proof. |
| 5 | Marketplace operator and agencies | Amazon at very high, 6,250. Plus Competitor 13, Competitor 33, Competitor 24, Competitor 25, Competitor 28, Competitor 30, Competitor 29, Competitor 31, Competitor 32. | Non buyers counted as prospects. |
| 6 | Wrong person linked | Two identified records carry LinkedIn URLs belonging to different individuals. | Outreach reaches the wrong person. |
| 7 | Industry mislabelling | Enterprise Account 53, ICP Brand 92 and Enterprise Account 80 all labelled Automotive. ICP Brand 80, the ICP Brand 89 brand, labelled Building materials. | Any cut using website enrichment needs rebuilding. |
| 8 | Non US entities present | Enterprise Account 37, Enterprise Account 32, ICP Brand 62, Enterprise Account 41 India, despite 100% US delivery. | Likely global parent resolution. Confirm before reporting reach. |
| 9 | Model disagreement | Account model reports 830 identified people, prospect model reports 1,815, same period. | Pick one definition before either is reported. |
All nine bias account intelligence upward, which means every ABM figure is biased upward until these are excluded. Exclusion list in section 06.
The clearest evidence that the targeting work landed, and the clearest evidence of what it cost. Comparing the top 30 engaged accounts by clicks in each month.
7,274 impressions / 264 clicks / 2 leads / 18 of 30 at 10,001+ employees
Both June leads came from 10,001+ enterprises: Enterprise Account 13 and Enterprise Account 62. Real leads, from companies that will not buy a marketplace growth retainer.
2,999 impressions / 101 clicks / 0 leads / 0 of 30 at 10,001+ employees
Nineteen of the thirty are true ICP consumer, beauty and wellness brands in the 11 to 1,000 employee range. Only one account, the agency Competitor 33, persisted from June's top 30.
Enterprise presence in the top engaged cohort went from 18 of 30 to zero in one month, replaced by exactly the DTC beauty, wellness and consumer brands the strategy targets. This is a defensible, named, verifiable result of the exclusion work.
Clicks in the top cohort fell from 264 to 101, impressions from 7,274 to 2,999, and leads from 2 to zero. The quality improved and the volume collapsed. Present both halves together or the first half will not survive scrutiny.
Click any column to sort. Dual signal means the account engaged through both LinkedIn ads and the website, which is the strongest available intent marker before deal stages are populated.
Honest tiering. The blazing hot tier is small because the evidence supports a small tier, and inflating it is how an ABM programme loses sales team trust.
2 accounts / immediate outreach
Do: verify both LinkedIn profiles by hand, then a named referral or founder to founder approach from the CEO. Do not route these through a generic sequence.
19 accounts / the July ICP cohort
Do: build a dedicated matched audience from this list, run the wellness and cosmetics creative that already carries the best rates, and put a bottom funnel offer behind it. None of these has produced a lead yet, so this is a conversion test, not a retargeting formality.
4 accounts / signal present, fit unconfirmed
Do: confirm buying authority and whether marketplace management is already in house before any spend is directed here.
13+ named entities / apply before next list build
Do: add to the exclusion list alongside the enterprise work already under way, then rebuild the engaged account counts. Expect the headline account numbers to fall once applied.
Where accounts stop moving, what it costs, and the specific fix. Dollar figures are estimated from impression share where noted.
| Bottleneck | Evidence | Cost | Fix |
|---|---|---|---|
| Impression hoarding by the low tier | 240 accounts absorbed 75,358 impressions, 34.5% of company resolved delivery, and 2,218 clicks, for 1 lead. 314 impressions per account against 18 at very high. | ~$8,427 estimated |
Frequency cap plus exclusion. Redirect toward the 482 very high accounts already producing 11 of 17 leads. |
| Reach without engagement | 7,231 accounts, 73.9% of everything reached, sit at very low engagement with 89,836 impressions and 198 clicks between them. Zero leads. | ~$10,046 estimated |
Tighten to the matched ABM list. Broad reach is producing volume, not accounts. |
| Deal stages unpopulated | The DemandSense-to-HubSpot sync joins every ad-touched company, but deal stage, amount and close date are blank on 417 of 418 records. | Blocks the revenue view |
Populate deal stage, amount and close date in HubSpot; DemandSense surfaces them automatically. |
| Website identity is not the ICP | 1,325 companies and 1,662 visits, but the highest engaging named people are a professor, a nurse, a massage therapist, an interpreter, a teacher and a school aide. | Corrupts scoring |
Exclude careers paths from account scoring, then rebuild. The careers page is the likely driver. |
| Self, competitor and agency contamination | Nine documented defects including the client's own company ranking first and a direct competitor at very high engagement. | Inflates every ABM figure |
Apply the exclusion list in section 06 and restate the counts. |
| Named list not safe to hand over | Two of forty identified people carry LinkedIn URLs for different individuals. Ten of forty have no company resolved at all. | Outreach risk |
Manual verification pass on any name before it reaches a sales sequence. |
Six moves, ordered. Two are gated on the client populating HubSpot deal stages and are marked rather than presented as ready.
Remove both Northwind Commerce entities, Impactable, Amazon and the ten agency and competitor accounts. Rebuild engaged account counts and tell the client the numbers moved and why, before the next report shows a drop.
240 accounts are taking a third of delivery for one lead. Frequency cap and exclude, then push the recovered budget into the 482 very high engagement accounts converting at 2.89%.
Nineteen genuine ICP brands surfaced in July. Load them as a matched audience, run the cosmetics and wellness creative that already performs, and put a real bottom funnel offer behind it.
See section 02b. Joined on email domain rather than company name, because only 2 of 17 LinkedIn labelled leads carry a company name. 17 click verified, 28 impression and click verified, 82 ecosystem. Lead with the 17 and the ICP Brand 108 card.
Not a company size problem as the audience analysis guessed. Ten of seventeen arrived with a personal email address and no company name. Six resolve to real brands: ICP Brand 99 twice, ICP Brand 78, ICP Brand 79, ICP Brand 108, ICP Brand 67. The defensible qualified count is closer to 6 than 17.
Exclude careers traffic, rebuild account scores, and reconcile the 830 against 1,815 people discrepancy. Until then the website layer adds noise to the ABM picture rather than intent.