By FroggyAds. Published 18 July 2026; updated 15 October 2026. We operate the FroggyAds self-serve advertising platform.
Your advertising account, website analytics and order system each describe a different part of the purchase. Bring them together without losing those distinctions. We help you use campaign and source evidence to make more informed traffic purchases, while your own business records establish which responses became useful leads or orders.
Quick answer: Advertising analytics connects recorded spending and delivery with visitor actions and confirmed business outcomes. The aim is a defensible next decision: continue a source, investigate a tracking gap or repair the destination. Start by matching the same campaign and period across your records. Then check what remains unmatched before treating a dashboard total as a complete explanation.
The Google Analytics: Get started with advertising documentation describes that provider’s reporting environment. The workflow here applies the broader reconciliation principle: understand what each record represents before combining it with another system’s numbers.
Advertising analytics connects the media purchase with the customer journey and the final business record. Each source contributes a different kind of evidence.
The buying account tells you about its recorded spend and delivery. Website analytics describes observed actions under its collection settings. Your order or lead system records the business status you accept. None of those roles makes the other records redundant. A useful analysis connects them while leaving their boundaries visible.
Start with the question you need to answer, such as whether a source deserves another allocation. You may need only a campaign export, a small event extract and an order-status summary. Build that connection carefully before adding more channels, dimensions or dashboards. The smallest reliable comparison often reveals more than a large report with unexplained totals.
A platform click, a measured visit and a confirmed order are different observations. Some clicks may not produce a measured visit. A visitor can return later, and an order can change status after its initial confirmation. Differences between the totals therefore deserve investigation, but they do not automatically establish a traffic or reporting fault.
Begin with the definitions and the reporting dates. A report grouped by click date can answer a different question from one grouped by purchase date. A business system may include all orders while an analytics report includes only attributed orders. Write down the boundary before deciding which total should control a buying change.
FroggyAds gives you a practical way to respond once the difference is understood. Use supported source, targeting and spending controls to refine the media purchase. If the evidence instead points to an incomplete destination or tracking setup, fix that part first. Better analysis directs effort to the place that can actually change the outcome.
Choose FroggyAds when you want direct campaign control alongside a clear record of what you purchased. Start with one offer and a question your destination can answer. Keep the relevant campaign references in your measurement setup so a useful finding can lead to a source, audience, bid or budget adjustment.
Create My Free Account
| # | Connection | Practical check |
|---|---|---|
| 1 | Campaign identity | Match a stable campaign reference, not just a display name that may have changed or been reused. |
| 2 | Reporting day | Check time zones and whether the date describes delivery, the customer action or the latest status update. |
| 3 | Recorded spending | Confirm currency and the media-cost boundary. Keep account deposits and balances outside the campaign-spend total. |
| 4 | Source reference | Retain the identifier your buying setup actually supplies. Leave an unknown reference unassigned rather than inventing a publisher. |
| 5 | Creative and destination | Keep the version that ran during the recorded period. A current screenshot cannot explain an older visitor experience. |
| 6 | Event and order identity | Distinguish repeated event messages from separate business outcomes. Test the actual deduplication rule. |
| 7 | Business status | Add acceptance, rejection, cancellation or refund information from the responsible system without deleting the history. |
| 8 | Combination result | Compare row counts and spending totals before and after matching. Explain additions, exclusions and unresolved differences. |
For a Google Analytics implementation, Google Analytics’ traffic-source guidance explains how manually tagged destination URLs provide campaign dimensions. Follow the current instructions for the tool you use. Confirm that your actual FroggyAds campaign and tracker pass the intended references; a parameter name in a planning document does not prove that the live journey preserved its value.
State the action the analysis will support. You might need to choose between sources or establish whether a checkout repair changed completed orders. Keep the question narrow enough that the required records are identifiable. A broad request to explain all performance invites a report that is large but difficult to use.
Choose the campaign dates and the date on which you will check business status. Retain both. Customers may finish later than they click, so a fresh period may still be incomplete. Label a preliminary result clearly and plan a later readback instead of silently comparing it with a settled period.
Keep the files or snapshots used in the review, with their extraction time and filters. Work on copies when correcting names or formats. If a colleague asks why yesterday’s total differs, you can examine the actual inputs instead of relying on a dashboard that has already updated.
Check whether spending appears once per campaign day or once per creative, source or device. Do the same for events and orders. This reveals when a single spend row would match many event rows. Decide how to summarize the records before combining them, and retain the original totals for comparison.
Bring dates, currencies and campaign identifiers onto a stated comparison basis. Keep a small mapping for renamed campaigns or equivalent labels. Do not remove distinctions merely to make a match succeed. A mapping should explain a real relationship that the team can verify, not conceal a missing connection.
Use the appropriate order or event identity to identify repeated messages. A returning customer can place a second legitimate order, so a customer identifier alone is not enough. Keep the duplicate rule explicit and inspect examples. The aim is to count each intended business outcome once, without discarding valid activity.
Create a visible category for events without a usable campaign reference and for orders still awaiting acceptance. Check whether the missing records cluster around a particular destination version or period. Their pattern can identify a repair. Hiding them would make the matched portion appear more complete than the evidence supports.
Check media spend, distinct outcomes and known exclusions against their original records. Investigate any unexpected increase after a join. Confirm whether a filter removed a valid campaign day or business outcome. Only then calculate the comparison that will inform spending, with the remaining limitations beside it.
A missing source reference calls for a tracking investigation. A working route with poor accepted outcomes may justify a media adjustment. Name the action and the evidence behind it. FroggyAds controls help you implement the buying part; your website and business systems supply their own corrective steps.
Preserve the date of the adjustment and compare the later records on the same basis. Check whether the original discrepancy disappeared and whether business outcomes improved. If the apparent gain depends on a different definition or incomplete period, explain that before presenting the change as a successful optimization.
Our Push, Native, Display and Pop options let you test different ways of introducing an offer. Keep the format and creative reference with the destination history where your setup supports it. That makes later analysis useful: you can distinguish the invitation people received from the page and business process they encountered afterward.
Create My Free Account
Use each record for the fact it can establish. Reconcile the relationships between records instead of choosing whichever report shows the largest result.
| Question | Useful starting record | What to check before acting |
|---|---|---|
| What media did we purchase? | The buying account’s recorded campaign spend and delivery. | Period, currency, filters and whether the total changed during later processing. |
| What did measured visitors do? | The analytics event record for the relevant destination. | Collection settings, event meaning, redirects and any known missing period. |
| How many distinct orders were created? | Unique order records in the commerce system. | Test orders, repeated messages and whether the same identifier was reused. |
| Which outcomes can the business accept? | The responsible order or lead-status record. | Acceptance criteria, cancellations and unresolved cases at the review date. |
| Which source can receive credit? | The supported campaign references and stated attribution method. | Missing references, competing credit claims and differences in attribution scope. |
| What changed between two periods? | The campaign, creative and destination change history. | Timing, other simultaneous changes and whether each period had comparable follow-up. |
A reporting system becomes easier to maintain when every important input has a clear role. The table below separates those roles. It also shows what to inspect when two reports appear to disagree. An unresolved difference can limit a conclusion without making every other observation unusable.
Inspect a small set of purchase events against the order references. A confirmation reload or repeated message may explain the surplus, but verify the actual cause. Correct the counting implementation before rewarding the campaign for extra sales that the order system cannot substantiate. Keep the valid order history and the original event extract for comparison.
The form count may be accurate while the commercial interpretation is incomplete. Match submissions to qualification status and preserve reasons for rejection. Location fit, service eligibility or unreachable contact details can suggest different changes. Use the pattern to choose between a clearer offer, revised audience or improved follow-up process.
Keep the single business outcome separate from each channel’s attribution report. Both systems may assign credit under their own rules. Adding their credited conversions would not produce a unique-customer total. Explain the competing views and use a consistent decision method before moving budget between channels.
A campaign-day cost can appear beside every matched order if the relationship is handled incorrectly. Summing that repeated column inflates spending without another media purchase. Compare the total before and after combining records. Summarize at a compatible level or use an appropriate model before calculating campaign economics.
A report containing only matched rows can conceal valid orders or paid delivery. Keep the unmatched set and explain why it is excluded from a particular comparison. Investigate the loss of references before deciding that one source produced no value. Missing attribution and missing business outcomes are different problems.
A renamed campaign can split one history into two labels or merge unrelated activity. Retain stable identifiers and a dated naming map. Keep creative and destination versions separately. This lets you review the offer that actually ran instead of attributing older results to the current configuration.
An attribution model distributes credit; a normal report does not establish what would have happened without the advertising. Avoid calling attributed orders incremental sales unless a suitable causal study supports that claim. Use the available records to describe association and buying economics while keeping the causal question distinct.
A preliminary order can later be cancelled or rejected. Keep the original campaign period connected to the later status where your process supports it. Show when the record was reviewed. Otherwise, recent campaigns can look strong simply because the business has not yet finished checking their outcomes.
The practical standard is reproducibility. Another person should be able to follow the key inputs and understand why the conclusion is limited or actionable. That does not require a complicated data platform. It requires a clear relationship between the records and a visible explanation of the exceptions.
Use FroggyAds to apply the media decision you can support: refine an audience, adjust a bid or restrict an unsuitable source where those controls are available. Keep the original comparison and the reason for the adjustment. Your next analysis can then examine the effect of a concrete change rather than an unexplained increase in activity.
Create My Free Account
A small campaign does not need an elaborate reporting stack to answer a focused question. A carefully checked export and an order-status summary may be enough. The important cost is the work required to keep the connection trustworthy: preparing inputs, resolving exceptions and revisiting outcomes when their business status changes.
Automation becomes valuable when the same reliable task repeats often enough to justify it. First establish the correct matching and summarization rules on a manageable sample. Then automate that known process with checks for unexpected row counts, missing fields and altered totals. A faster incorrect join only spreads the same error across more reports.
Choose additional detail when it changes a real decision. If a source comparison is actionable, preserve the references needed for it. If a dimension cannot be interpreted or used responsibly, collecting more of it may add maintenance without helping the purchase. Put the saved effort into checking the offer and improving the next controlled campaign.
We provide self-serve buying controls for supported formats, audiences, bids, budgets and sources. Those controls make analysis operational: you can translate a verified finding into the next traffic purchase. Preserve the campaign and source information your actual account and tracker expose, and connect it with the business outcome at the level your implementation supports.
If you use Google Analytics, Google Analytics’ campaign-data import guidance explains matching external campaign data with collected activity. This is guidance for that provider, not a promise of a native FroggyAds connector. Confirm the current available export or tracking method and test the fields you intend to use before depending on an automated report.
Your business system remains responsible for its order or lead status. FroggyAds does not automatically know every refund, cancellation or qualification decision. A configured tracker, pixel or postback can support measurement, but the meaning of its event still matters. Keep the integration understandable enough that a missing update can be investigated.
A vendor demonstration is most useful when it includes an awkward record, a missing reference and a later status change. These cases reveal more than a polished chart built from perfect sample data. Choose the tool whose behavior your team can explain and maintain, with access limited to the information needed for the job.
When the totals diverge, trace one known order through the journey. Check its campaign reference, event time and final business status. Find the first point at which the records disagree. This can distinguish a lost parameter from a duplicate event or a later cancellation without requiring you to guess from the overall conversion rate.
Google Analytics’ transaction-ID guidance documents purchase deduplication for web streams, with specific requirements for transaction identifiers. Apply the current instructions for your own implementation; do not assume every stream or tool behaves the same way. Keep customer-identifying information out of advertising URL parameters and use appropriate access for records needed in the investigation.
After correcting the issue, repeat the relevant controlled check and note which reporting periods were affected. Keep the earlier uncertainty visible if the missing data cannot be recovered. You can still make a limited decision from reliable records while arranging the next campaign to collect the evidence that was absent.
Metrics define measurements such as click rate or accepted-order cost. Advertising analytics connects the records needed to interpret those measurements and make a buying decision. This guide focuses on campaign identity, matching, duplicates, business status and unresolved differences. Use the metric definition and the underlying records together.
No single architecture suits every advertiser. A small, well-defined comparison may work with controlled exports and an order-status summary. Start with the decision and required records. Add automation or storage when the repeated work and complexity justify it, while keeping the original inputs and checks available.
They describe different observations and may use different collection conditions. Check the destination, redirects, reporting period and analytics implementation. A difference deserves investigation, but the totals should not be forced to agree without an explanation. Keep any known missing period or measurement limitation beside the result.
Use the references that your actual campaign, tracker and business system support. Stable campaign and order identifiers are usually more useful than editable display names. Verify the matching rule on known records and keep unmatched outcomes visible. Do not infer personal identity from incomplete or probabilistic information.
A single cost row may match several event or order rows. If the cost is repeated on each resulting row and then summed, the total becomes inflated. Compare spending before and after the combination, and summarize or model the relationship correctly before calculating acquisition cost.
Keep valid business outcomes in the appropriate overall business total, while showing that their source is unresolved. Do not distribute them across sources without a defensible method. The account-level result may remain useful even when a source-level comparison needs repair or more complete references.
No. Attribution assigns credit under a stated method. Incrementality asks what happened because of advertising compared with what would otherwise have happened. A routine matched report does not answer that causal question on its own. Describe the evidence you have without turning credited orders into a stronger claim.
Retain the original action and add the later business status. Show the acceptance rule and the date of review. This helps you distinguish response volume from usable outcomes and avoids rewarding a campaign only because its early records have not yet been checked.
Do not assume that a particular native connector or automatic synchronization is available. Verify the current tracking, export and integration options for your setup. Our campaign controls support media buying; your analytics and business systems provide their own records and require a correctly configured connection where you use one.
Use it when the relevant inputs reconcile under clear rules and the remaining uncertainty does not undermine that decision. Explain the proposed adjustment and what you will inspect next. FroggyAds lets you act through supported source, audience, bid and budget controls while retaining a clear reason for the purchase.
These existing references cover Google reporting terminology and broader publishing or accessibility standards. Use the relevant product documentation for the product you actually configure. The nearby citations above support specific implementation distinctions. None of these sources establishes a guaranteed FroggyAds result or replaces a check of your own collected records.
A colleague should be able to understand the decision without rebuilding every chart. Keep these eight items together with the final comparison so the next review starts from the same evidence.
State the buying or destination change you recommend, together with the question it answers. A source restriction and a tracking repair need different evidence. Keep that distinction in the handover so implementation does not turn an unresolved data problem into an unnecessary media change.
Include campaign dates, extraction time and the date on which business status was checked. Mark preliminary results as preliminary. This allows the next reviewer to understand whether newly appearing orders reflect fresh performance or simply a later read of the same purchase period.
Preserve spending and distinct-outcome totals before combining files. Include the filters that produced them. These figures give the next reviewer a practical control: if the assembled report changes, they can identify whether the difference began in the source or in the transformation.
Explain which fields connect the records and what one row means at the final reporting level. State how repeated matches are handled. Plain language is enough for a simple comparison, provided another person can follow why each cost and outcome is counted once.
List the missing references, delayed statuses or excluded records that could change the recommendation. Describe their size and affected period where known. An exception list is useful when it tells the next person what to investigate; an unexplained warning symbol is not a substitute.
Identify the business status used for the final result and how rejected or cancelled activity appears. Keep the earlier response count available for diagnosis. A colleague should be able to distinguish a revision to customer quality from a change in the meaning of an accepted outcome.
Record the advertisement, destination and buying changes that could explain a difference. Include when they took effect. A source exposed to two different offers should not be discussed as if nothing changed. This history prevents a convenient but unsupported explanation from becoming the next campaign plan.
Finish with the specific observation that would confirm, weaken or change the recommendation. It might be a repaired source reference, completed lead review or a later order-status readback. Agree how that evidence will reach the buyer so the analysis leads to a timely operational decision.
Turn a checked finding into a focused campaign adjustment. Our self-serve controls let you manage the supported audience, format, spending and sources while your own records explain the customer result. Start with a manageable purchase and keep the next question clear enough that its answer can improve what you do afterward.
Create My Free AccountConsider a hypothetical campaign with USD 600 of media spend. Its raw event extract contains 30 purchase messages, representing 20 distinct orders. All 20 orders are matched to this campaign. Five are later cancelled, leaving 15 accepted orders. These invented figures illustrate reconciliation; they are not FroggyAds prices, customer results or a prediction.
The reporting-tool choice is a separate task covered by the existing Marketing Analytics Tools resource. Here, the immediate job is to reconcile the same campaign period. Keep the original spending total and all four outcome counts available so a reviewer can see how the final denominator was reached.
Using all 30 messages would suggest USD 20 per purchase message. Using 20 distinct orders gives USD 30 per order created. Using 15 accepted orders gives USD 40 per accepted order. Each division is arithmetically valid, but only its correctly labelled result answers the intended business question.
| Record stage | Count in this hypothetical example | What the next check establishes |
|---|---|---|
| Raw purchase messages | 30 messages | Inspect repeated messages; this is not yet a distinct-order count. |
| Distinct orders created | 20 orders | Ten excess messages are excluded from the distinct-order total. |
| Accepted business outcomes | 15 orders | Five cancellations are outside the accepted count; other business costs remain separate. |
| Supported source assignment | 12 assigned; 3 unassigned | Source-level conclusions retain the unresolved attribution gap. |
There is a separate joining trap. If the single USD 600 campaign-spend row is repeated beside all 20 order rows, summing that column produces USD 12,000. No additional media was bought. Preserve the USD 600 source total and combine the data at a compatible reporting level before using the result.
Suppose 12 accepted orders have a supported source reference and three do not. All 15 remain accepted business outcomes for this example, but a source comparison must disclose the three unmatched orders. Do not award them to the apparent winning source or erase them to make the totals align.
Use FroggyAds to act on the part of the finding you can support. Investigate missing references before claiming a source-level winner, and retain the USD 40 account-level accepted-order cost as a properly labelled observation. It does not establish profit or incremental sales. Create your free FroggyAds account
Your first useful result may be a better explanation of a disagreement. Use that explanation to decide what to repair, what can be compared and which campaign adjustment is justified.
