Pass Around
A pass around puts one and the same work product in front of several reviewers who read it separately, each at a time that suits them, with no session and no facilitator. BABOK counts it among its seven review types and defines it in one sentence: an informal technique in which multiple reviewers provide verbal or written feedback, on a common copy of the work product or by passing it from one person to the next (§10.37.3.2). The number of reviewers is what goes in; out come n sets of comments that overlap, repeat one another and sometimes contradict one another, with nobody to reconcile them but the author, alone and after the fact. A pass around surfaces the disagreements; it does not settle them. That is its own particular use: several independent readers are the only instrument that shows that a document reads two ways.
Goal
A pass around puts several independent reviewers in front of the same text and reports what each of them read in it. A single reading has nothing to compare against: it reports what one reviewer understood and says nothing about the question that decides the fate of a glossary, a specification or a business rule. Does this document mean the same thing to two people who did not write it?
BABOK lists seven objectives for a review (§10.37.3.1), and a pass around serves three of them. Remove defects: several reviewers find more of them than one does, and they do not find the same ones. Ensure conformance to specifications or standards: each reviewer brings their own, those of their site, their function or their mandate. Educate reviewers about the work product: circulating a document is how a distributed team ends up sharing it, and Wiegers credits the pass around with that objective while he withholds it from the desk check (10.37e) (Table 3-3). To those add an objective BABOK's list does not carry: testing the readability of the document, in the strict sense of finding out whether it reads the same way twice.
Four of the seven objectives are out of its reach. Establish consensus on an approach or a solution: Wiegers reserves this objective for the reviews that hold a session and withholds it from the pass around (Table 3-3). The technique collects the positions, and a consensus is built in a room, with a formal walkthrough (10.37b). Ensure the work product is complete and correct: completeness is measured against what ought to be there, and a reviewer working alone sees what is written. Wiegers illustrates this on his own project plan: the pass around did not surface the tasks missing from the work breakdown structure, which the team meeting found afterwards. An inspection (10.37a) reaches completeness because it consolidates the readings in a session. Measure work product quality presupposes defect metrics, so a log, so an inspection. Answer a question, resolve an issue or explore alternatives: BABOK attaches this objective to work in process (§10.37.2), so to an ad hoc review (10.37g) or an informal walkthrough (10.37d), where one peer is asked and one answer comes back. A pass around puts a frozen deliverable in front of n independent readers: it poses the question and n answers come back; bringing them together remains the author's work.
The deliverable is the collated comments, some of which contradict each other, and then what the author has reconciled. What is left is whatever the author could not reconcile. That remainder is an output of the technique: the conflicts that survive are what gets escalated, with both positions intact. The defect log belongs to the reviews that hold a consolidation session, and the decision to those that have a room.
Usage
When to use it
- Several perspectives are wanted on one artefact: legal, business, IT, another site, another language.
- The reviewers cannot be assembled: geographical distance or calendars; each contributes at a time that suits them (§10.37.4.1).
- More reviewers than a room holds: Wiegers had each chapter of his own book reviewed by sixteen people.
- The question is whether the document reads the same way twice: only independent readers settle it.
- A term, a rule or a standard has to hold from one site to the next: the conflicts that come back are the finding.
- Ahead of an expensive review: a pass around removes the easy errors quickly, and the inspection that follows recoups its cost only on the defects that remain.
- No record is required: the output is a corrected work product.
When not to use it
- A record of the review is required (audit, quality gate, approval): the number of reviewers attests to nothing. An inspection (10.37a) logs the defects and verifies the rework; a single issue review (10.37c) attests conformance to one named standard.
- The perspectives must be reconciled: an author alone does not arbitrate five opinions. A formal walkthrough (10.37b) reconciles them in session, with a facilitator who is not the author.
- One competent reviewer is enough: on a low-risk artefact, collation is pure cost. A desk check (10.37e) costs less, and Wiegers reserves control of review cost for it (Table 3-3).
Description
BABOK gives this technique no steps, and the technique asks for only two: announce a window, freeze a version. Everything else sits in a decision taken before the pass around starts, often without anyone noticing, and in the collation work that comes afterwards.
Choosing how the work product circulates
BABOK names two modes in its definition: the work product may be reviewed on a common copy or passed from one person to the next (§10.37.3.2). A third one exists, and most organisations practise it: separate copies, sent the same day to five people who never see each other. BABOK does not count it among the modes. It does describe a cost among its limitations, without naming either the mode or the technique: if review comments are shared and discussed via e-mail there may be many messages to process, which makes it difficult for the author to resolve disagreements or differences in suggested changes (§10.37.4.2). Attaching that limitation to the separate copies is a reading.
The limitation can only arise where several reviewers comment separately, outside any session. A desk check (10.37e) has one reviewer, so there is no second reading to contradict the first. An inspection (10.37a), a formal walkthrough (10.37b) and a single issue review (10.37c) hold a session, and the consolidation happens in the room. BABOK gives a facilitator and a scribe to the first two and judges a facilitator helpful for the third (Table 10.37.1). An informal walkthrough (10.37d) holds a session too, and the same table gives it a scribe. In an ad hoc review (10.37g), the business analyst seeks informal review or assistance from a peer (§10.37.3.2): one opinion comes back, so there is no disagreement to reconcile. Of BABOK's seven reviews, the pass around is the only one whose definition calls for several reviewers and provides for no session: the many-messages limitation can be aimed at nothing else.
The three modes differ on exactly one point, and everything else follows from it: does a reviewer see what the others wrote?
Separate copies. Each reviewer reads the text as it stands, without knowing what the others saw in it. Independence is at its maximum, so coverage is too: five readers who have not influenced each other report five different readings. They also report the same obvious defect five times, and their disagreements land raw on the author's desk. This mode costs the most to collate, and you get it by attaching a document to an e-mail.
The common copy. One copy, in a shared space, annotated by everyone for a set period. Wiegers describes this mode precisely and says what it buys: every reviewer sees the comments already written, which reduces redundancy and reveals differences of interpretation between reviewers. The collation almost does itself. The price is independence: someone who reads the others' remarks before writing their own is no longer reading the document alone, and BABOK already notes that the author may find it difficult to establish that an independent review was done by each participant (§10.37.4.2). In this mode, independence degrades with nothing to signal it.
The serial pass. One copy, one reviewer at a time, each inheriting the previous one's annotations. The elapsed time becomes the sum of the readings: five reviewers at three days each make a fifteen-day review. That elapsed time is why this mode stays rare. It converges, and it places a specialist last, on an already improved text. It also propagates the first reviewer's faulty correction to all the others, who read the error before they read the text.
Independence buys coverage and costs collation. The author pays for the collation in every case, and the mode only decides whether they pay during the pass around or after it. The tool you reach for chooses the mode, whether or not you notice.
What a pass around requires
Two things are enough. An announced window: Wiegers collects comments for a set period that closes the review. Without it, a reviewer gives an opinion on a version that no longer exists, and Wiegers observes that late contributors waste their time on an obsolete work product. A frozen version: if the reviewers are not all reading the same text, the collation means nothing. Whatever is added to those two points is ceremony.
Collating
The comments come back unsorted. The author deduplicates them, then separates the preferences from the conflicts. That second operation decides the quality of the review.
A preference is two reviewers wanting two wordings. The author decides. A conflict is two reviewers holding two incompatible definitions of the same term. The author cannot decide, because the question is which of the two definitions the organisation actually applies. Reconcile what is reconcilable and escalate the rest with both positions intact: that is where the technique delivers what it promises.
A finding that has to survive the review, be tracked and be closed has left the pass around and belongs to item tracking (10.26).
Where the pass around sits
BABOK lists its seven review types in descending order of formality and calls the first three formal techniques, the next four informal techniques (§10.37.3.2). Two binary questions cut across that ordering: is the session planned; does the reviewer examine the work product alone before speaking? The pass around answers no, then yes. The desk check (10.37e) answers the same way. Those two questions locate the two techniques without separating them.
Wiegers, who supplies this spectrum, describes the pass around as a desk check carried out by several people at once, and he is right about what goes in. His Table 3-3 puts the review objectives on one axis and the methods on the other. That is where the two techniques stop coinciding: the pass around gains the education of the reviewers and the assessment of readability, which the desk check does not get, and it loses control of review cost, which the desk check keeps. The same act repeated by several people becomes a different instrument, because the comments that come back have to be reconciled and there is nobody to do it but the author. The number is the input; the reconciliation is the output, the one the desk check does not produce.
The traps
Taking the number for assurance. "Five people reviewed it" reads as five times the confidence a single reviewer would give it, and BABOK says the opposite: for desk checks and pass arounds alike, it may be difficult for the author to establish that an independent review was done by each participant (§10.37.4.2). The number makes the problem worse, because it makes doubt uncomfortable. This trap does real damage: it cancels an inspection (10.37a) the team believes it has held.
The invisible reviewer. Wiegers gives it as the technique's ordinary failure mode: of those who volunteered, several never answer. With a single reviewer, silence is visible and blocks the review. With five it is no longer visible: comments come back and the review "took place". The two people who never opened the document are the ones whose perspective the author wanted. The remedy is a list: know who answered and say so.
Reconciling by picking a side. Two incompatible definitions come back, and the author keeps one and publishes. The conflict is still there. It is no longer written down anywhere, and the organisation goes on working with two definitions of which only one is documented. The review had produced its finding, and the author erased it.
The debate in the margin. On a common copy, two reviewers answer each other in the comments. Wiegers warns against those exchanges and tells the two reviewers to settle the point in a direct conversation. Two reviewers debating in an annotation thread have convened a session with no facilitator, and the author inherits it.
Letting the review become a session. The day the reviewers are brought together to go through the comments, a formal walkthrough (10.37b) is under way. It is often the right answer, and it deserves to be a decision: the session has a facilitator, a scribe and a cost.
AI considerations
The difficulty particular to this technique is the collation: n sets of comments to be brought together against a text and against each other. It is a text-comparison job, and a large language model does it well.
Deduplicating. Five reviewers, one defect, five reports written five ways. A model clusters them in seconds, and it does it as well on the thirtieth occurrence as on the first.
Surfacing the contradictions. This is the high-value use, because the author misses these collisions more easily than the rest: two comments that are reasonable separately and incompatible together. The author checks each comment against the document, one by one, whereas a model reads the n sets in one pass and sees the collision.
Drafting the reconciliation as a proposal, marked as one, for the person who has to decide.
A model does not add a reviewer. A pass around collects positions inside the organisation: a site, a function, a mandate, a stake. A model occupies none of them: added to the pass around, it brings text where the technique asks for a perspective.
It does not reconcile a conflict, because a conflict is settled by an authority. When Geneva and Zurich return two incompatible definitions of the same term, the question is who decides. A model that picks one has filed the disagreement away where nobody will find it again.
Evidentiary weight degrades as the summary improves. A clean summary produced by a model from five sets of comments looks like a review record, while BABOK's limitation still applies to every one of the five reviewers. The document appears to attest to what it does not attest to, and it is harder to challenge than a disorderly e-mail thread.
Data protection. Circulating a work product is already a distribution question, and putting it through a model is a second one. A glossary or a specification carries internal terminology, client data and unpublished rules. The revised FADP, the Swiss Federal Act on Data Protection, raises a question that is settled before the productivity one. The organisation fixes its rule before the pass around starts: which model, which data, hosted where.
Examples
This example shows what a pass around produces: collated comments carrying a conflict, then a reconciliation that is partly an escalation. The glossary entry was circulated as separate copies, sent the same day to five reviewers across two sites.
| Entry submitted | Active client: "A client who has carried out at least one transaction in the past twelve months." | |
|---|---|---|
| Reviewer | Site | Comment returned |
| Compliance | Geneva | "Twelve months: matches our practice. Make it explicit that the transaction is initiated by the client." |
| Reporting | Geneva | "Same remark: initiated by the client." |
| Wealth management | Zurich | "Active here means: a custody account opened and not closed. A client holding CHF 800'000 in custody with no transaction is active for us." |
| Data | Zurich | "Which number does the monthly report deliver? Today it counts the open custody accounts, not the transactions." |
| Legal | Geneva | No reply. |
| Reconciled | Active client (transactional): "A client who has carried out at least one transaction on their own initiative in the past twelve months." |
|---|---|
| Reconciled | Relationship client (holdings): "A client holding at least one account or custody account that is open and not terminated." |
| Not settled, escalated | Which of the two definitions feeds the "active clients" indicator of the monthly reporting. Both coexist in the bank today, and the choice does not belong to the author of the glossary. |
The author reconciled what was reconcilable: the redundancy of the two Geneva remarks and the ambiguity about who initiates the transaction. They could not reconcile the conflict, because the question is which of the two definitions the bank publishes in its reporting. A desk check (10.37e) would not have produced this: one reviewer returns one definition, and the author publishes it. The conflict is the finding, and it takes several independent readers to produce it.
The circulation mode can be read off the same artefact. On a common copy the second Geneva remark would never have been written, and the Zurich reviewer would have read the Geneva definition before drafting their own, which they might not have drafted at all.
Visualisations
The three circulation modes form a spatial object: an axis, the visibility between reviewers, three topologies placed on it and two quantities that reach their maximum at the same end, independence and collation cost. A table would give the four properties and lose what holds them together: the order of the three modes, the topology that produces their properties and the fact that the serial pass occupies the middle of the axis while carrying the minimum of independence. Choosing the mode is the technique's only real decision.
BABOK's formality spectrum places the pass around among its six neighbours. The choice between the seven reviews belongs to reviews (10.37).
The artefact actually produced is a set of collated comments with the reconciliation. It goes into a table, because a set of comments is a list of rows read by scanning.
Cost
| Phase | Level | Justification |
|---|---|---|
| Preparation | Low | Wiegers counts no planning in this technique (Table 3-1). The author picks the reviewers, freezes a version, announces an end date. Minutes. |
| Execution | Low to medium | Several readings instead of one, but they run in parallel and each reviewer chooses their moment (§10.37.4.1). The effort is real and it is borne by the reviewers. In serial mode the elapsed time is the sum of the readings. |
| Documentation | Medium | Collating the comment sets, deduplicating, separating the preferences from the conflicts and reconciling what can be reconciled is real work, and it falls entirely on the author. This is the price of the number: Wiegers grants control of review cost to the desk check and withholds it from the pass around (Table 3-3). |
Tooling
The common copy needs a shared document where the comments and the suggestions are visible to everybody: a SharePoint or M365 space, a Confluence page, a review thread on a merge request. The set period Wiegers speaks of is structural in this mode. Without it, the document stays open to annotation until nobody knows which version they are reading.
The serial pass needs routing: an approval workflow or a mail chain with one recipient at a time.
Separate copies need an attachment and five recipients.
A finding that has to be tracked after the review closes belongs to item tracking (10.26), which is a different technique and has tools of its own.
Sources
- IIBA, A Guide to the Business Analysis Body of Knowledge (BABOK Guide) v3, §10.37 Reviews.
- Karl E. Wiegers, Peer Reviews in Software: A Practical Guide, Addison-Wesley, 2002, chapter 3, "Peer Review Formality Spectrum" (Table 3-1, activities by review type; Table 3-3, objectives and methods).