Things Keep Falling Through the Cracks? Find Where First

Key Takeaway

When things keep falling through the cracks at work, the cracks are not random and they are rarely about effort. They sit in a few predictable places: handoffs with no one on the receiving end, work with no trigger to start it, recurring tasks and renewals nobody owns, deliverables nobody tracks, and exceptions handled from memory. The usual advice is to buy a task app, but most dropped work never became a task for an app to track. So find where first: take the last ten things that slipped, trace each back to the step it fell from, and tally them by crack type. Then fix the structure at that step with a named receiver, a trigger, a definition of done, and a single place of record. Only then decide whether a tool should enforce it. Across 16 businesses we studied, only 27% of the work was documented, so most cracks sit in steps nobody has written down.

Why Things Keep Falling Through the Cracks

Nobody drops work on purpose. The proposal that never went out, the renewal that lapsed: each had someone who meant to handle it.

Owners on discovery calls often describe a manual process this way: it's workable, it works, though things can slip through the cracks. That is accurate. A manual process works on a normal week, with the usual people in their usual seats. The cracks open on the other weeks: when someone is out, volume spikes, or a new hire takes over a step nobody explained.

That is why the usual fixes disappoint. Telling the team to be more careful assumes the problem is attention. Replacing the person assumes the problem is the person, but the substitution test in system problem or people problem usually says otherwise: a capable replacement in the same seat misses the same thing, because the miss is built into the work. Buying a task app assumes the problem is tracking, which we come back to below.

The real cause is structural, and it shows up in the numbers. When we gap-analyzed 16 small businesses across 68 roles and 461 process areas, the average business had 27% of its work documented. In 50.3% of role areas there was no documentation at all, and 82% of teams were below 50% coverage. Undocumented work is not necessarily broken. But the spaces between undocumented steps, where one person's part ends and the next begins, are where nobody has decided who catches what. Those spaces are the cracks.

The Five Places Things Fall Through the Cracks

One question does more diagnostic work than any other on a discovery call: where does it usually fall? Owners can almost always point at the spot, and the answers sort into five types of crack.

1. Handoffs with no receiver

Work passes from one person or team to another, and nobody on the receiving side owns it once it lands. Sales closes a deal and posts in a group chat. Someone forwards a request to "the team." The sender believes the work was handed off, and everyone on the receiving side assumes someone else has it. A message sent to a group is not a handoff. It is a hope. A swimlane diagram shows these fastest, because every arrow crossing a lane line is a handoff that needs a catcher.

2. Work with no trigger

A step that is supposed to happen, with nothing that makes it happen. No event, date, or status change says it is time, so it runs when somebody remembers. One owner described crews starting work before anyone had sent a proposal or opened a purchase order. The paperwork had no trigger of its own, so the crew's start date became the real trigger and the paperwork chased it. We walk through that fix in our business process improvement examples. The same pattern covers the follow-up after a quote and the invoice after a job closes.

3. Recurring work and renewals nobody owns

Insurance and license renewals, maintenance agreements, annual service visits, vendor contracts. They are rare enough that nobody builds a habit around them and important enough that missing one is expensive. They usually live in one person's calendar or one person's head. When that person leaves, the renewal does not leave a gap you can see. It leaves a surprise months later.

4. Deliverables nobody tracks

The work got done, but the output has no defined home: reports, photos, signed forms, inspection notes. One business needed a service report from about three months earlier and nobody could find it. It existed. It was simply wherever its author happened to save it. The miss is not the work but the record of the work, and it surfaces later as a dispute, a repeat visit, or an afternoon of searching.

5. Exceptions handled by memory

The standard path is understood. The exceptions are not: the rush job, the customer with special billing terms, the emergency call. The veteran handles them without thinking, so nobody notices they are different until the veteran is out. The tell is that misses cluster around unusual cases and one person's absence. Writing them down is the work of documenting decision points, and the frequency test for process map exceptions decides which ones belong on the map.

The Crack Types at a Glance

The Crack Finder: Trace Your Last 10 Misses

Memory is a poor locator. Owners remember the most painful miss, not the most common one. The crack finder replaces that impression with a count, and it takes an afternoon.

1. List the last ten things that slipped. Late callbacks, missed renewals, jobs started without approval, files nobody could find. Write each as one sentence about what happened, not who did it. If you cannot reach ten, keep an interruption log for a week. Every interruption that starts with "did anyone ever" is a miss worth listing.
2. Trace each one back. Find the last step where the work was definitely on track and the first step where it was not. The crack sits between them. Ask the people involved where it usually falls.
3. Name the step and the crack type. If a miss fits two types, pick the one closest to where it fell.
4. Tally. Count misses by crack type and by location. If several misses land on the same step, that step is a crack, not a coincidence.
5. Pick one location. Start with the step that collected the most misses, or the one where a single miss costs the most.

The worksheet needs five columns. A few filled-in rows look like this:

Keep the trace about steps, not people. And separate lost work from slow work: if the work arrived late because it sat in a queue, that is a bottleneck, and finding the bottleneck in your business is the better tool.

Fix the Structure Before You Pick a Tool

Once you know where things fall, the fix is usually one of four structural changes, and none of them requires software. They are the Improve stage from our guide to business process improvement, aimed at one step.

A named receiver. Every handoff gets one person, not a team, who owns the work once it lands, and a defined way they learn it has arrived. Name a second person who notices when the handoff did not happen. That split between doing the work and catching the miss is the difference between responsible and accountable, and the catch is the part most businesses skip.

A trigger. Every step that matters starts from something specific: a signed proposal, a status change, a date, a completed previous step. "When it's ready" is not a trigger. "When the purchase order is open" is. For recurring work, the trigger is a date with lead time built in, so the owner hears about the renewal weeks ahead, not on the day it lapses.

A definition of done. Write down what finished looks like, including the record. A job is not done when the work is done. It is done when the report is filed, the photos are attached, and the invoice is triggered.

A single place of record. Each kind of work and output gets one official home. Not the group chat and the inbox and a shared drive. One. That is the idea behind a single source of truth, and it is what makes a three-month-old service report findable in a minute.

The fifth crack, exceptions, closes with the written decision rule from the table, because a rule on paper does not take vacation.

Will a Task App Stop Things Falling Through the Cracks?

Most advice on this problem ends with a product recommendation: a task manager, a project management tool, a shared board. Those tools are good at tracking tasks someone created. The trouble is that most dropped work never became a task. The handoff that slipped was a chat message. The renewal was a date in one person's head. A task app has nothing to track until someone decides what the task is, who owns it, and what starts it.

That is why a new app can help for a few weeks and then fade. The team fills it with the tasks they already remember, which were never the ones falling through.

A tool earns its place after the structure is fixed, when it enforces a rule that already exists:

• Recurring dates. Once every renewal is on one list with an owner, reminders are exactly what software does well.
• Required fields and gates. A job management tool can block scheduling until a proposal is accepted, or refuse to close a job without a report attached, but only if someone decided those were the rules.
• Queues and volume. When handoffs or recurring items outgrow what a person can track, a queue with a named owner beats a chat thread, and automation keeps the trigger reliable.

This is the Solve stage of business process improvement: software runs a process, it does not create one. Fix the step, then let the tool hold it in place.

How to Keep the Cracks Closed

Cracks reopen. A new hire takes over a step with no handoff notes. New software adds a lane nobody mapped. Growth turns a handoff that worked at ten jobs a week into one that strains at forty.

• Trace every new miss. When something slips, find the step before finding the person, and add it to your tally.
• Re-run the crack finder every quarter. If the misses cluster somewhere new, that is your next fix. If they still cluster in the old spot, the fix did not hold.
• Review handoffs when anything changes. A new hire, role, tool, or service line creates new handoffs. Check each for a receiver and a trigger before it goes live.
• Update the record when the rule changes. A fix that lives only in a conversation drifts back. Write it into the documented process using the change loop in keeping procedures current.
• Count the misses. Misses per month, by crack type, tells you whether the cracks are closing. Record it before you change anything, and check it after.

Where to Start

Write down the last ten things that slipped, trace each to its step, and count. Fix the step with the most misses using a receiver, a trigger, a definition of done, or a single place of record. Run it for a few weeks, count again, and only then decide whether a tool should hold it in place. It is the same order laid out in our business process improvement guide: document, diagnose, improve, then solve.

Frequently Asked Questions

Why do things keep falling through the cracks at work?

Because the work crosses gaps nobody owns. Most dropped work falls from a few predictable places: handoffs with no named receiver, steps with no trigger, recurring tasks and renewals that live in one person's head, deliverables with no defined home, and exceptions handled from memory. It is rarely about effort: a capable replacement in the same seat would usually miss the same thing.

How do you stop things falling through the cracks?

Find where they fall first. List the last ten things that slipped, trace each back to the step it fell from, and count them by crack type. Then fix the structure at the step with the most misses: one named receiver, a trigger, a definition of done, and a single place of record. Add a tool only after that, to enforce the rule you have written.

What is a dropped handoff?

A dropped handoff is work passed from one person, team, or system to another that never gets picked up. It usually happens because the handoff had no named receiver: the work was posted to a group, sent to a shared inbox, or mentioned in a meeting, and everyone on the receiving side assumed someone else had it. The fix is one named receiver per handoff and a clear signal that the work has arrived.

Will a task management app stop tasks falling through the cracks?

Not on its own. A task app tracks tasks someone created, and most dropped work was never turned into a task. Once the steps where things fall have owners, triggers, and a definition of done, a tool is very good at enforcing them, especially for recurring dates, required fields, and high-volume handoffs.

Is it a people problem when things keep slipping?

Usually not. Ask whether a capable new person in the same seat would miss the same thing for the same reason. If they would, the problem is the structure: a missing receiver, trigger, or record, and coaching the person changes nothing because the next person walks into the same gap. If the miss is specific to one person in a well-defined step, it is a people problem, but that is rarer than it feels in the moment.