Small-business outsourcing research

Which tasks do small-business owners keep after delegating admin work?

A protocol for listing the tasks an owner still performs after a delegation, classifying them, and testing whether the boundary matches the written role.

Research question and scope

This protocol asks which tasks a small-business owner still performs after delegating administrative work, and whether those tasks match the written boundary. The unit is one owner-assistant pair inside the observation window. The study lists the tasks the owner keeps, the tasks the assistant performs, and the tasks that both touch. It does not evaluate whether the owner is good at delegating or whether the assistant is effective.

Set the scope to one small business, one owner, and one assistant role. Record the role brief, the access list, and the task list in force at the start. If the role changes during the window, keep the earlier definition for the full sample or analyze the change separately. A new task can look like a boundary problem when it is only a new service.

Population and unit of analysis

Define the population as all recurring administrative tasks that the business performed during the window, including tasks the owner kept and tasks the assistant performed. The unit is the task, identified by a short name and the system it uses. A task that is performed weekly is one unit with a frequency, not one unit per occurrence. A one-off project is outside the population unless it becomes recurring.

Record the task category, the system, the frequency, and whether it involves money, customer data, access, or a legal obligation. These strata matter because a task that touches money or access carries a different risk from a task that formats a document. Include tasks that are shared between the owner and the assistant, because shared tasks are where boundaries blur. Test the unit definition on two ordinary weeks before collection begins.

Definitions that make the measure reproducible

Define an owner-kept task as one the owner performs without the assistant's involvement in the decision or the action. Define a delegated task as one the assistant performs end to end under the written brief. Define a shared task as one where the assistant prepares and the owner decides or sends. Define a boundary mismatch as a task whose actual handling differs from the written role brief.

Do not count a task as delegated when the owner still checks every output before it leaves. That is a shared task with a review step, even if the assistant does the work. Do not count a task as owner-kept when the assistant prepares it and the owner only signs. The classification should follow the work, not the job title. Each definition should include one positive example and one boundary example.

Evidence standard and data dictionary

For each task, capture the name, the category, the system, the frequency, the person who performs it, and the person who decides. Add the review step, the access required, and whether the task touches money, customer data, or a legal obligation. Record the role brief line that covers the task, if any. Code each field as present, missing, or not applicable.

Also record whether the owner has been asked to approve an exception in the last month and how it was handled. These fields help explain a boundary that looks clear on paper but is blurred in practice. If the task is performed in a personal inbox, record that fact and the reason. A second coder should be able to classify the task from the interview and the records.

Sampling and collection

Start with the calendar, the inbox, and the task board rather than with a list of tasks the owner remembers. Record the period, the sources, and the selection rule. Ask the owner and the assistant to describe a normal week separately, then compare the two accounts. The difference between the accounts is often the most useful evidence.

Collect the role brief, the access list, and a sample of completed work. Do not reconstruct a task from a single conversation. If a task is seasonal, record the season and the last time it was performed. Note the collection date separately from the period so the reader can see how much of the record has aged.

Analysis and reporting

Report the number of tasks by category, the share the owner keeps, the share delegated, and the share shared. Show the counts by risk category and by frequency. Present the share of tasks with a written brief line and the share with a boundary mismatch. Use the same denominator for every rate and show the number of tasks with incomplete evidence.

Inspect the mismatches and describe the pattern without naming the people. A cluster of money-related tasks in the owner-kept group may be appropriate, while a cluster of access tasks in the shared group may be a gap. Compare the mismatch with the exceptions handled in the last month, because an exception is where the boundary is tested. Report the tasks that were delegated without the required access.

Limitations and sources of error

The record cannot show whether the owner enjoys a task or whether the assistant is ready for more. It can show only who performs the work and who decides. A boundary that looks inefficient may be a deliberate risk choice, and a boundary that looks clear may hide an unwritten review. The protocol also cannot capture the tasks that are never written down because they happen in passing.

Calendars and inboxes can omit verbal requests, interruptions, and work done on a personal device. A task that is performed only when the owner is away may not appear in the sample. Record those cases and their treatment instead of dropping them silently. If the owner and the assistant describe the same task differently, record both accounts rather than forcing one.

Review cadence and next action

Give the result to the owner and the assistant and ask them to choose one observable change. The change might be a written brief line for a shared task, an access adjustment, or a new review step for a money-related task. Record the owner, the effective date, and the measure that will show whether the change worked. Avoid changing the brief and the access list at the same time.

Repeat the protocol with the same definitions after the change and compare the same fields. If mismatches fell but the owner's workload did not, the tasks may have moved to a different shared category. If access improved but review did not, the assistant may be doing more without a check. Close the review when the owner accepts the finding and the next audit is scheduled.

Sources

Frequently asked questions

Does a favorable sample prove the routine caused the result?

No. The design is observational and supports a bounded operating decision, not a universal causal claim.

What should an owner review first?

Start with high-consequence defects, missing denominators, exceptions, and cases that required the owner to reconstruct the evidence.

Related research