# Know what generated the impressions and clicks you sell.

Advertisers pay for people seeing ads. Attach actor evidence to each ad event and see which publishers and placements deserve a closer look before you bill, credit or renew.

For ad networks and platforms serving ads through their own tags.

Status: Discuss an integration

## Every automated impression is a billing question.

An impression is served, maybe rendered, maybe seen. When a bot or crawler generates it, the advertiser may still be billed, and the dispute lands on your team weeks later.

Evidence on the event itself lets you separate supply that needs review from supply that is fine, alongside the filters you already run. Classification alone does not decide billability; your contract does.

## Where $200,000 of billed impressions went

- Impressions served: 100M
- Billed spend: $200,000
- Confirmed ineligible, beyond existing filters: 2M (Only after review against your contract terms.)
- Spend at issue: $4,000

Illustrative example. 100M impressions at a $2 CPM; 2% confirmed ineligible beyond existing filters. Hypothetical.

## The inventory report, two ways.

| | Reported | With evidence |
|---|---|---|
| Billable impressions | 100M | 98M |
| Spend billed | $200,000 | $196,000 |
| Advertiser credits at risk | $0 | $4,000 |

Illustrative example.

## Your numbers

Billed spend = impressions ÷ 1,000 × CPM. Spend at issue = billed spend × share confirmed ineligible.

## One impression, followed to the bill.

- Publisher: pub-2291
- Placement: sidebar-300x250
- Actor evidence: bot · confidence high
- Billing review: Flagged for supplier review

Your existing filters missed it. Your team decides whether it is billable.

## How it fits

1. **Observe on your tag.** Run actor evidence in your approved ad tag or SDK, where the impression happens.
2. **Carry the event id.** Attach the evidence to the impression or click id in your event ledger.
3. **Join billing.** Compare with served, rendered and viewable counts and your existing filters.
4. **Decide by contract.** Review suppliers, credit advertisers or keep billing, under your own terms.

## What your team can do

- **Review a publisher.** See classified, unknown and confirmed events by supply source.
- **Answer a complaint.** Show the evidence behind a disputed campaign.
- **Compare with your filters.** Measure overlap, additions and disagreements before any enforcement.

## Before you start

### Is this an accredited ad-verification product?

No. It is a proposed integration that adds actor evidence on your own tags. It does not replace accredited measurement.

### Can a landing-page script see these impressions?

No. Impressions that never lead to a visit need the evidence in your ad tag or SDK. Sandboxes, consent and blocked scripts limit browser coverage.

### Does a bot label prove fraud?

No. Legitimate crawlers can still be excluded from billable counts, and classification alone does not decide billability.

### Will it block bids before they happen?

We are not promising pre-bid blocking, refunds or coverage inside closed ad platforms. A pilot starts with reporting alongside your existing controls.

[Discuss an ad-platform pilot](mailto:support@doubleagent.so?subject=Use%20case%3A%20Ad%20traffic%20quality)

- [Double Agent Expose](https://doubleagent.so/)
- [Conversion quality](https://doubleagent.so/use-cases/conversion-quality/)
- [All use cases](https://doubleagent.so/use-cases/)