Reviewed: 9 October 2026. The order ledger, channel allocations and costs below are original fictional exercises. They are not an actual analytics account, Google's algorithm, campaign results or evidence of causal marketing impact.
Marketing analytics asks what a measurement means, whether its underlying records reconcile and which decision the evidence can support. A report showing channel revenue is not automatically a profit report. Fractional credit is not necessarily a count of customers, and a change in credit allocation is not a change in completed sales.
For a business in India, begin with a defined period, transaction status, currency and metric. Then separate operational records from analytics attribution. This guide develops a complete reconciliation so you can see where a channel comparison is useful and where it exceeds the supplied evidence.
Define the question and metric scope
Our fictional business wants to understand how two editorial credit-allocation exercises describe three paid orders in one period. Every monetary amount below is in rupees. The exercise uses a paid-order ledger, with no tax or refund amount supplied. It does not define these figures as GA4 Purchase revenue or a financial-accounting revenue policy.
The question is: “Do the allocated totals reconcile to the specified paid-order amounts, and how does channel credit differ?” It is not: “How many additional orders did advertising cause?” No experiment or comparison group is supplied to answer that second question.
Google's attribution report assigns key-event and revenue credit to touchpoints and compares model results. Its documentation also distinguishes reporting time bases and missing dimension values. Those definitions explain why scope matters; they do not validate this fictional ledger or its invented allocations. Google Analytics attribution report.
Keep preparing
Continue with Sarkari Resume Templates₹299 — coaching के एक महीने से काफ़ी सस्ता / far cheaper than a month of coachingReconcile transaction status first
The complete original transaction packet is:
| Item | Details |
|---|---|
| O1 | Specified amount: ₹1,000; Supplied status: Paid; Included in this paid-order exercise?: Yes |
| O2 | Specified amount: ₹2,000; Supplied status: Paid; Included in this paid-order exercise?: Yes |
| O3 | Specified amount: ₹1,000; Supplied status: Paid; Included in this paid-order exercise?: Yes |
| O4 | Specified amount: ₹1,500; Supplied status: Cancelled before payment; Included in this paid-order exercise?: No |
The included amount is ₹1,000 + ₹2,000 + ₹1,000 = ₹4,000 across three orders. Adding O4 would incorrectly produce ₹5,500 for the metric defined here. A cancelled order may still matter operationally, but it belongs in a separate status measure.
The packet does not identify unique customers. Three orders could belong to three people, one person or another combination. Do not label the order count “three customers” without a suitable identity record. Likewise, the packet supplies no new-versus-returning classification.
Inspect the first credit allocation
Allocation A is an invented editorial arithmetic example. Each order receives one total unit of order credit, distributed as follows:
| Item | Details |
|---|---|
| O1 | Search share: 1; Email share: 0; Unknown-source share: 0; Total share: 1 |
| O2 | Search share: 0.5; Email share: 0.5; Unknown-source share: 0; Total share: 1 |
| O3 | Search share: 0; Email share: 0; Unknown-source share: 1; Total share: 1 |
Apply each share to that order's specified amount. O1 assigns ₹1,000 to search. O2 assigns ₹1,000 to search and ₹1,000 to email. O3 leaves ₹1,000 in the unknown-source bucket.
| Item | Details |
|---|---|
| Search | Allocated order credit: 1.5; Allocated amount: ₹2,000 |
| Allocated order credit: 0.5; Allocated amount: ₹1,000 | |
| Unknown source | Allocated order credit: 1; Allocated amount: ₹1,000 |
| Total | Allocated order credit: 3; Allocated amount: ₹4,000 |
The 1.5 search credit is not one and a half people. It is the sum of shares under this exercise. The unknown amount is not zero and should not be dropped to make the named channels appear more complete.
Compare a second allocation without changing the orders
Allocation B keeps O1 and O3 unchanged. For O2, it assigns 0.25 to search and 0.75 to email. These shares still total one.
For O2's ₹2,000, that means ₹500 search credit and ₹1,500 email credit. The resulting totals are:
| Item | Details |
|---|---|
| Search | A order credit: 1.5; B order credit: 1.25; A amount: ₹2,000; B amount: ₹1,500 |
| A order credit: 0.5; B order credit: 0.75; A amount: ₹1,000; B amount: ₹1,500 | |
| Unknown source | A order credit: 1; B order credit: 1; A amount: ₹1,000; B amount: ₹1,000 |
| Total | A order credit: 3; B order credit: 3; A amount: ₹4,000; B amount: ₹4,000 |
Search amount credit decreases by ₹500, or 25% relative to its ₹2,000 under A. Email amount credit increases by ₹500, or 50% relative to its ₹1,000 under A. Actual orders and their specified paid amounts remain unchanged.
The differences answer an allocation question. They do not show search losing sales or email producing extra sales. A and B are neither named GA4 selectable models nor approximations of Google's data-driven method.
Preserve unknown information honestly
O3 has no supplied source evidence. The exercise leaves its credit in an editorial “unknown source” bucket. Do not rename it organic search because that would improve a presentation, or assume a missing source means no marketing touchpoint existed.
Google documents several report placeholders for missing or unavailable dimensions, including “not set” and “Unassigned.” Their meanings differ. Our invented bucket is not a recreation of one of those specific GA4 classifications; investigating a real report requires its actual dimension and configuration.
Also avoid teaching obsolete model selectors as current features. Google's report documentation says first-click, linear, time-decay and position-based attribution models have been unavailable since November 2023. The editorial shares here are for arithmetic practice, not instructions to select those models in GA4.
Compare credit and cost within their limits
The exercise supplies ₹800 in search advertising cost and ₹300 in email delivery cost for the same period. Under A, search allocated amount divided by its specified cost is ₹2,000/₹800 = 2.5. Email's equivalent is ₹1,000/₹300, approximately 3.33.
Under B, those ratios become ₹1,500/₹800 = 1.875 and ₹1,500/₹300 = 5. The costs did not change; only the assigned amount did. Label the ratios as allocated amount divided by specified channel cost. They are not proven incremental returns, profit or a basis for promising a campaign outcome.
The total included amount minus the two specified costs is ₹4,000 − ₹800 − ₹300 = ₹2,900. It cannot be called profit because product costs, staff time, fees, refunds and tax treatment are not supplied. The unknown-source bucket also has no supplied cost. Missing cost is not evidence that a channel was free.
Test two changes to the evidence
First, suppose email cost is unavailable in a second exercise variant. You can still compare its allocated amount across A and B, but you cannot calculate its amount-to-cost ratio. Do not retain the earlier ₹300 as though it were confirmed in the new variant, and do not divide by zero to represent a missing value.
Second, suppose O3 receives a verified partial refund of ₹200 in a later period. The original period's supplied packet contained no refund. Before changing reports, specify whether you are reviewing the original period's paid-order amount, a later net-amount measure or a restated period under an established policy. Under a separately defined all-three-orders net-amount exercise, ₹4,000 − ₹200 = ₹3,800. That number is not automatically the value a particular analytics report will display.
These variants show why date and metric definitions belong in the analysis. A correct subtraction can still answer the wrong question if scope is unclear.
Write a decision note the evidence supports
Here is the complete original analysis note:
Scope: three specified paid orders in the defined period, totalling ₹4,000; cancelled O4 excluded. Allocation A and B each preserve three units of order credit and ₹4,000 in amount credit. B shifts ₹500 from search to email while leaving ₹1,000 unknown. Specified search and email costs are ₹800 and ₹300. Their ratios depend on allocation and do not establish profit or causal lift. Before changing spend, investigate missing source information, cost completeness, comparable reporting periods and the evidence needed for the business decision. No unique-customer count or incremental sales effect is established.
The note gives a manager both the comparison and its limits. It does not disguise uncertainty as an instruction to move all spending to the larger ratio.
Build practical analytics skills from the exercise
Recreate the order table in a spreadsheet or another tool you know. Check row totals, status exclusions, allocation shares and bucket totals. Keep the original records separate from calculated outputs so a reviewer can trace a number back to its inputs.
A learning portfolio should describe the exercise as fictional. It can demonstrate reconciliation and careful metric interpretation, but not experience running a client campaign or an actual GA4 property. If you use SQL or a dashboard, claim that work only after completing and checking it.
In a real role, the employer's data access, privacy requirements and measurement setup matter. Do not download customer records or change attribution settings merely to reproduce this example. The article has accessed no analytics account.
Frequently asked questions
Is attributed revenue proof that a channel caused a sale?
No. Attribution assigns credit under a method. Causal incrementality needs additional evidence appropriate to that question.
Can fractional credit be described as customers?
Not from this record. Orders, unique customers and allocated shares are separate metrics.
Should unknown-source rows be excluded?
Preserve them when needed to reconcile the defined total. Investigate missing information rather than silently assigning it elsewhere.
For connecting measurement to a business choice, see strategic marketing planning. For campaign-level definitions and a separate session exercise, see digital marketing strategy.