Three suppliers can use the same headline number for different products. One may promise clicks, another sessions, and a third a mixed-country delivery. Compare scope before price. The fictional examples below use 10,000 advertised visits and show a clause, an unknown, and a clarification question.
1. Purchased unit: clicks, sessions or unique visitors
The first problem is that a click is an ad interaction, a session is a visit recorded after a page loads, and a unique visitor is an attempt to count people rather than visits. These units can produce very different totals. For example, Quote A says “10,000 clicks,” Quote B says “10,000 sessions,” and Quote C says “10,000 visitors,” but none of those phrases proves that all three buyers receive the same thing.
Analytics documentation defines a session as a period of interaction that can end after inactivity, while total users are unique users who trigger an event in a selected period. Consent, tagging, browsers, and setup also affect the measurement. A practical step is to compare the supplier’s delivery event with your own reporting.
- Sample clause: Quote A: “10,000 clicks to the supplied URL.” Quote B: “10,000 sessions, measured by our system.” Quote C: “10,000 unique visitors.”
- Unknown: Whether a click that fails to load, a repeat visit, or a blocked analytics event counts.
- Clarification question: “Which event ends delivery, how do you count repeats, and what evidence lets me reconcile your number with my analytics?”
You may want to record the purchased unit in the worksheet as click, session, or unique visitor. It is better not to assign a quality score while the unit remains unclear.
2. Ad source and placement format
The same unit can have a different value if the source and placement differ. A display placement on a relevant page, a recommendation tile, and a pop-under can all send a person to your URL, yet they create different expectations about attention and user experience. A supplier that says only “real advertising” has left the most important delivery detail open.
You can review the scope with the website traffic service details in mind: ask what type of placement is suitable for your page, then request the actual inventory description in writing. A source list, placement format, and landing-page behavior give you something concrete to review instead of relying on a label.
- Sample clause: Quote A: “Traffic from display placements.” Quote B: “Traffic from content recommendations.” Quote C: “Mixed ad inventory.”
- Unknown: The host-page categories, ad position, creative size, redirect behavior, and whether the placement can be changed.
- Clarification question: “What format and host-page categories will deliver the traffic, and can you provide a placement description or sample before launch?”
A useful check is to ask whether the ad opens the landing page directly and whether the supplier records source and placement identifiers. You can clear inventory makes later performance review possible.
3. Location targeting and its limits
A country label does not prove that every visit comes from that country. Routing, shared networks, travel, and location-detection differences affect geographic reports. Quote A says “United States,” Quote B says “North America,” and Quote C says “worldwide,” so the 10,000 visits are not equal coverage.
- Sample clause: Quote A: “US targeting.” Quote B: “North America, subject to availability.” Quote C: “Global delivery.”
- Unknown: The country or region split, permitted variance, and treatment of traffic that cannot be located confidently.
- Clarification question: “Which countries and regions are included, what percentage range applies to each, and how do you handle unclassified locations?”
Put the requested locations beside the delivered locations in your worksheet. A simple variance is delivered visits - requested visits; a negative value shows a shortfall, while a positive value does not prove that the extra visits are useful.
4. Device mix
Device mix changes how a landing page behaves. Mobile-heavy delivery can expose slow forms or small buttons, while desktop-heavy delivery can misrepresent a mobile audience. “Responsive” is not a device split.
- Sample clause: Quote A: “All devices.” Quote B: “70% mobile, 30% desktop.” Quote C: “Mobile-first delivery.”
- Unknown: Whether tablet traffic is separate, how the split is measured, and whether the supplier can pace a requested range.
- Clarification question: “What device percentages do you expect, are tablets reported separately, and what variance is allowed?”
It helps to use the device split as a delivery check, not as a promise about conversions. If your page has a form, review a small sample of mobile and desktop sessions for page load, form visibility, and error events.
5. Exclusions and duplicate or invalid-traffic handling
A traffic total is easier to review when the supplier explains what it removes. Repeated requests, crawlers, data-center activity, and refreshes can inflate a count. Missing signals do not automatically prove that a visit is invalid.
Recognized measurement guidance separates routine invalid-traffic filtering from harder detection cases and warns that privacy limits can reduce available signals. A practical step is to ask for the supplier’s exclusion rules and its method for avoiding false positives. A claim that every unclassified visit is “human” is too broad.
- Sample clause: Quote A: “Bot-filtered traffic.” Quote B: “Clean visits.” Quote C: “No exclusions stated.”
- Unknown: Duplicate window, crawler rules, data-center treatment, post-delivery filtering, and whether exclusions are included in the headline total.
- Clarification question: “Which exclusions apply before and after delivery, how are duplicates handled, and what summary can you provide without exposing personal data?”
You may want to mark an unanswered exclusion question as unknown. It is better not to convert it into a zero or a passing score. The goal is a documented scope, not a made-up certainty.
6. Delivery window and pacing
Delivery timing affects what your analytics can tell you. Ten thousand visits in two hours may create a sharp spike, while the same total spread across ten days can support a steadier page and audience review. “Fast delivery” is not a schedule.
- Sample clause: Quote A: “Delivered within 48 hours.” Quote B: “Seven-day campaign.” Quote C: “Start immediately, timing varies.”
- Unknown: Start trigger, daily limits, pauses, weekend behavior, and what happens if inventory is temporarily unavailable.
- Clarification question: “What are the start and end conditions, expected daily range, and pause or restart rules?”
For the fictional 10,000-visit example, a seven-day window has an average of 10,000 / 7 = 1,428.57 visits per day. That is a planning average, not a guarantee. A supplier that states a daily range lets you compare pacing without treating the average as an exact promise.
7. Reporting access and supporting evidence
A final total without supporting fields makes it difficult to explain a change in results. You need to know when visits arrived, which source or placement sent them, and how location and device were classified. Your own analytics may show fewer sessions because tagging, consent, or measurement methods differ.
- Sample clause: Quote A: “End report supplied.” Quote B: “Live dashboard included.” Quote C: “Total delivered on request.”
- Unknown: Export format, reporting delay, time zone, source fields, device fields, and whether the report is a delivery record or an analytics estimate.
- Clarification question: “Can I receive a row-level or daily summary with date, source, placement, country, device, clicks, and sessions, plus the reporting time zone?”
You can download the traffic package comparison worksheet and fill the reporting fields before you approve a quote. A blank cell is a useful finding because it identifies the next question.
8. Delivery review and shortfall terms
A supplier may define delivery as the number purchased, while you may define it as the number reaching a working page. If 9,600 of 10,000 clicks load, the gap needs a review method and remedy.
- Sample clause: Quote A: “Replacement for missed volume.” Quote B: “Best effort.” Quote C: “Delivery complete when sent.”
- Unknown: Review period, reconciliation source, minimum shortfall, replacement window, and exclusions for downtime or a changed landing page.
- Clarification question: “What counts as a shortfall, which records settle the count, and when do I need to report a discrepancy?”
It helps to keep your own URL uptime and campaign-change log. If you edit the destination during delivery, note the time because a broken or redirected page can make a traffic discrepancy harder to interpret.
9. Total price, fees and tax treatment
A comparison fails if one quote includes every charge and another adds setup, payment, reporting, or tax later. Quote A is $900, Quote B is $800 plus $100 setup, and Quote C is $850 before a hypothetical 5% tax. The totals are $900, $900, and $892.50. A practical step is to confirm the tax assumption rather than treating it as universal.
- Sample clause: Quote A: “$900 total.” Quote B: “$800 plus $100 setup.” Quote C: “$850, taxes extra if applicable.”
- Unknown: Currency conversion, payment fees, tax responsibility, refunds, and whether support or reporting costs extra.
- Clarification question: “What is the final payable amount in my currency, which fees or taxes may apply, and what happens if the agreed scope is not delivered?”
You may want to compare the final payable price per purchased unit only after the unit matches. In the example, $900 divided by 10,000 equals $0.09 per unit. That arithmetic is useful for scope-matched quotes, but it does not measure traffic quality.
A practical quote-review worksheet
You can copy the three fictional rows below into your review process, then replace each unknown with a written answer. The asset is a reusable CSV, and its unknown values are intentional placeholders rather than quality ratings.
| Quote | Unit | Source and format | Location | Device mix | Exclusions | Window | Reporting | Shortfall terms | Final price | Status |
|---|---|---|---|---|---|---|---|---|---|---|
| A | click | display placement | US, variance unknown | all devices | bots mentioned | 48 hours | end report | replacement mentioned | $900 | clarify |
| B | session | content recommendation | North America | 70% mobile | clean visits | 7 days | live dashboard | best effort | $900 | clarify |
| C | unique visitor | mixed inventory | global | mobile-first | unknown | timing varies | total on request | delivered when sent | $892.50* | clarify |
* Hypothetical calculation using $850 plus a hypothetical 5% tax. It helps to confirm the actual tax treatment with the supplier and the relevant jurisdiction.
A short email can keep the comparison factual:
Please confirm the purchased unit, ad source and placement format, location split and variance, device mix, exclusions, delivery window, reporting fields and time zone, shortfall remedy, and final payable price. Please also state which fields remain unknown and what records will support the final delivery count.
After scope is aligned, cost per visitor can help you assess price. For placement questions, see ad placement review checks.
Final check before you choose
A fair comparison has the same unit, defined source, location and device scope, exclusions, schedule, reporting, remedy, and final price. Leave missing fields as unknown and ask for clarification instead of inventing a score.
Traffic can support a measurement plan, but it cannot guarantee engagement, leads, or search visibility. Your decision becomes clearer when the supplier’s promise and your own check use the same words, dates, and counting method.