The task and original case
Reconcile overlapping conversion associations before reporting a campaign total. A conversion can appear under more than one channel in this invented case. The portfolio output should preserve those associations while also showing the number of unique conversion identities.
| Case element | Value or requirement |
|---|---|
| Inputs | Channel A is associated with 6 conversions, B with 8, with 3 shared conversion IDs. |
| Build | An attribution reconciliation with explicit conversion identities. |
This is an original invented exercise. It is independent practice, with no employer assessment or open-vacancy claim. Download the inputs, attempt your own result, then open the separate answer key.
Data and worked inputs
Build a result another person can follow
Use a conversion-to-channel table with one row per association. Channel A is associated with six conversions and B with eight; three conversion IDs occur in both. Create the distinct ID set for each channel, then calculate their intersection and union. Keep the association table so the overlap is inspectable.
pandas join documentation helps identify shared keys and unmatched membership. Declare whether duplicate rows within one channel are repeated records or distinct events before counting. This exercise uses unique conversion IDs for conversions and a separate count of channel associations. It does not select a credit-allocation model or claim which channel caused a conversion.
Sources[1]
Check the worked result
Unique conversions total 6 +8 −3 =11. Channel-associated counts sum to 14 because the three shared conversions each have two associations. Both 14 associations and 11 conversions can appear in the report when clearly labelled.
Do not add channel counts and describe 14 as unique conversions. Likewise, do not delete shared IDs from one channel merely to make channel totals sum to 11 unless you explicitly choose a credit rule and keep the original association evidence.
Make the failure visible
Duplicate an A–conversion row and verify that the unique conversion count stays unchanged under the declared identity rule. Add a third channel sharing two existing IDs; a two-set overlap subtraction is no longer a complete general solution. Use distinct IDs across all included rows, or calculate a fully declared attribution model instead.
Revise it and ask someone to reproduce it
Ask a peer to build the union from the actual IDs rather than trusting the headline counts. For a second revision, add one conversion linked only to B; unique conversions become 12 and associations 15. Preserve the overlap table and explain the difference between descriptive association and causal credit.
- The three shared IDs can be inspected.
- Fourteen associations remain distinct from eleven unique conversions.
- Duplicate associations and additional channels do not inflate unique identities.
Describe the work accurately
After completing your own version, you could adapt this project bullet: “Reconciled an invented channel-conversion ledger with explicit overlap, unique identity counts and separate attribution-association totals.”
Use Projects or Independent practice as the heading. Keep the input files, result and revision together. Describe only your own completed work; the invented numbers are model outputs, not employer results or measured business impact.