The symptom is familiar to every engineering lead: a pull request has been “in review” for six days, a security questionnaire went to the vendor and vanished, and the release everyone expected on Thursday is now a vague “soon”. Nothing is broken in the code. Work is simply parked behind other people’s replies. The diagnosis is almost always the same, and so is the fix: a waiting-for list with named owners and nudge dates.
The problem: work stalls in handoffs, silently
Delivery pipelines are built from handoffs: author to reviewer, developer to security, team to vendor, engineer to change board. Each handoff is a promise that someone else will act. When the promise is verbal or buried in a chat thread, there is no mechanism that notices when it lapses. The author assumes the reviewer is on it. The reviewer assumes the author will ping if it matters. Both are busy. Days pass.
Teams usually respond by adding process: stricter service levels, more dashboards, another standup. Those can help, but they treat the group, not the individual who needs to know what they themselves are waiting on. The cheapest first repair sits at personal level, and it works even when the surrounding system is untidy.
Diagnosing it in your own week
Try this ten-minute exercise. Open your messages from the last fortnight and list every request you sent that required an answer from someone else. Beside each, mark whether you received it. You will almost certainly find two or three that never came back, and at least one you had forgotten entirely. That number is your baseline cost of not tracking, and it is usually higher than expected.
The fix: a list with four fields
Create a document with one line per outstanding request. Each line needs four fields.
- What: the specific artefact or decision, such as “firewall rule change approval” or “penetration test report”.
- Who: the owner of the reply, by role or team.
- Asked: the date you made the request.
- Nudge: the date you will follow up if nothing has arrived.
The Nudge column is where the value lives. It converts “I should chase that sometime” into a scheduled action. On the morning of that date, you ping. If the answer arrives earlier, you strike the line.
A concrete engineering example
Suppose a team is migrating a customer-facing service to a new database cluster. The lead has several dependencies outside her control: the platform group must provision the cluster, the security team must approve the network rules, the vendor must confirm a licence change, and the product owner must sign off a maintenance window. Her list looks like this on Monday:
- Cluster provisioning, platform group, asked Monday, nudge Wednesday.
- Network rule approval, security team, asked Monday, nudge Thursday.
- Licence confirmation, vendor account contact, asked Monday, nudge next Monday.
- Maintenance window sign-off, product owner, asked Monday, nudge Tuesday.
On Tuesday she nudges the product owner and gets the window approved. On Wednesday the platform group admits the ticket was assigned to someone on leave; it is reassigned before it costs a day. On Thursday, the security approval is still open and she escalates to the team’s queue owner. Without the list, the migration would have discovered those gaps one by one, at the worst possible moments.
When requirements are specified up front, fewer surprises appear later. A deep walkthrough such as this engineering blueprint for gathering web project requirements lists many of the external inputs a project needs, and each external input is a ready-made line for the register.
Handling vendors who go quiet
Vendor silence deserves a protocol. When you send a request, state the date you need a reply and why. Record a nudge date two or three working days after that, and a second one after another similar interval. If the vendor still does not respond, change channel: ring the account manager or send the request to a secondary contact. Document each attempt on the list. This accomplishes two things: it moves the issue forward, and it creates a clear record should procurement or management ask why the integration slipped.
If you are scoping a custom build, asking the right questions at the start reduces the number of later dependencies. A checklist like a set of pre-build questions for a bespoke web application can help you spot, in advance, which outside parties you will have to chase.
Handling stalled reviews
Code review delays are the most common internal case. Add each review request to your list with the reviewer’s name or the team queue, the date and a nudge a working day or two later. When the nudge comes, post a concise message: what changed, how large it is and when you need it merged. If reviewers keep slipping, bring the pattern to the team retrospective with data from your list, not vague frustration. A shared set of numbers, such as how long requests typically wait, makes a better conversation than anecdotes.
Keeping it lightweight
The trick is to resist over-engineering. A plain text file or a spreadsheet beats any custom tool you will not maintain. Keep one line per request, no sub-tasks, no priority matrices. Review the list first thing every morning, and again before you finish for the day. If a line grows complicated, move the detail into the original ticket and keep only a pointer.
Adapting it for an on-call and incident context
Incidents create their own waiting lists. During an outage, the incident commander is waiting on a log export from the cloud provider, a rollback confirmation from the deploy team and a status update from the customer-support desk. Writing those three lines on a shared incident document, with the person responsible and a time to re-ask measured in minutes rather than days, stops the call from circling. Afterwards, the same list becomes raw material for the review: which waits were longest, and why?
The principle is the same at every scale. Only the nudge interval changes: minutes during an incident, days for a change request, weeks for a procurement question. Keep the unit appropriate to the situation, and the habit transfers with no extra effort.
What to do when you are the one being chased
The same list teaches empathy. When a colleague sends you a dated, specific reminder, you can usually act on it quickly, because the question is clear and the deadline is known. Take a cue from that when your own inbox fills with nudges: reply promptly with at least a revised date, even if you cannot deliver yet. “I will have this by Friday” is far more useful to the person tracking you than silence, and it gives them a new date to record.
A note on life outside the pipeline
Engineers also run long personal projects. One that comes up often is buying a home, which involves lengthy waits on replies. A site called Propraahi says it provides property research material for home buyers in India, and anyone using something like it would probably keep a note of questions sent out and answers pending. The same four fields apply neatly.
Measuring whether it works
After a month, look at your closed lines. How many were chased before they became blockers? How often did a nudge produce an answer within a day? Did your release dates slip less often? You need no formal metrics; a rough count is enough. If you see no improvement, the usual culprits are nudge dates set too far out or a list that is not reviewed daily. Tighten both and try again for another fortnight.
Common failure patterns
- The list is private and nobody else sees the blockers. Share the relevant lines in standups so the team can help unblock.
- Nudges turn into nagging. Vary the channel, stay brief and keep the tone friendly.
- Items never close. If something is no longer needed, tell the owner and cross it off.
- Escalation is skipped. After two polite nudges, involve the owner’s manager or a neutral coordinator.
The overall lesson is that most delivery delays are not technical puzzles but communication gaps, and communication gaps respond well to modest, consistent tracking. Build the list, assign the dates and look at it every day. In a few weeks the stalls that used to surprise you will start showing up while there is still time to act.

