What a Library of Things Can Measure

A shared drill can help with several household projects without becoming several different products. That is the practical attraction of a library of things: people gain access to equipment when they need it, while the same collection serves successive borrowers. Describing that activity accurately is a separate task. Loans, users and visits each reveal something useful, but they do not count the same event.

Reading Library of Things, in the United Kingdom, reported 1,049 item loans over the preceding twelve months in its August 2026 impact report. It estimated 2,098 visits by allowing two per loan, while noting that people may collect several items together or visit without borrowing. The distinction matters: an activity record and an estimate built from it answer different questions.

Begin with the event being counted

An item loan records an object passing into a borrowing arrangement. A transaction may contain more than one item. A borrower may complete several transactions during the reporting period. Before adding the totals to a public report, name the event behind each number. A broad label such as participation can conceal these differences and make an ordinary increase in repeat use look like an increase in distinct people served.

The right unit depends on the question. Staff examining how often equipment circulates need item-level records. Someone organising a collection session needs to know how many customers are expected and what they are collecting. A report about reach needs an appropriately defined count of distinct borrowers. None of these measures is automatically superior; the problem arises when one is presented as evidence for another without the necessary connection.

Keep the period and definition together

Attach the reporting period to the measure, not just to the cover of a document. A twelve-month loan total and the current number of members may describe different populations. State whether the borrower count includes anyone who borrowed during the period or only accounts active at its end. This makes comparisons easier to interpret and avoids changing a definition silently when a new report is prepared.

Separate item movement from customer trips

A collection trip can move several objects at once. A person might borrow a drill and an accessory together, then return both during another visit. Counting each item as a separate collection and return produces a different figure from counting the person's actual journeys to the service. Both records can be useful, but the first is a model of item movements, not necessarily observed footfall.

Some visits also occur without a completed loan. A visitor may ask about availability, inspect the catalogue or discuss a future project. If a report intends to describe attendance, it needs a way to include the relevant visits rather than deriving every one from loan transactions. Conversely, attendance alone does not prove that equipment was borrowed or that a project was completed.

Follow a fictional week through the records

Consider an invented service that lends twelve items during a week. Eight customers collect them: four take two items each and four take one each. There are twelve item loans, eight collection transactions and eight distinct borrowers. These figures are all correct for the example, but each describes a different aspect of the same activity. None should replace another merely because it produces the largest headline.

Suppose every customer returns their items together on a later visit. Across collection and return, the example contains sixteen customer visits. Multiplying the twelve item loans by two would produce twenty-four item-related movements. That calculation is understandable, but it overstates the number of customer visits in this particular example. The difference arises from grouping items, not from an error in the loan total.

Now add three separate visits from people asking about future borrowing without taking anything away. Recorded attendance becomes nineteen visits, while completed item loans remain twelve. This does not mean the service has gained three new borrowers. It has recorded three additional enquiries. If the same person makes two of those enquiries, the number of distinct visitors is different again. Define the measure before interpreting the total.

The example does not predict the experience of any real service. Its purpose is to show why a public report should distinguish a count taken from records from a figure estimated through a multiplier. A simple model can be useful when direct observation is unavailable, provided that its assumptions are visible and the result is not relabelled as a measured fact.

Describe repeat use alongside reach

Repeat borrowing can indicate that an existing member finds the service useful. New borrowing can indicate that the collection is reaching someone it had not served before. These are different contributions. A growing loan total may reflect either or both, and the distinction matters when choosing the next improvement. Outreach and support for returning users address different parts of the service.

Use a consistent definition of first-time borrowing. A person's first transaction this year is not necessarily their first transaction with the organisation. If older records are unavailable, describe the narrower measure honestly. Similarly, an account created during the period does not establish that its holder borrowed anything. Registration, a first completed loan and a later return loan belong to separate stages.

A useful report can show both the number of distinct borrowers and the distribution of their activity. An average alone may hide a small group borrowing frequently and many people borrowing once. That is not automatically a problem. It is a prompt to understand how the collection is being used and whether the service's opening arrangements and available items fit different needs.

Ask about the alternative to borrowing

A loan does not automatically replace the purchase of a new product. Without the service, a person might have borrowed from a neighbour, hired commercially, bought second-hand, delayed the task or decided not to undertake it. These possibilities lead to different interpretations. Counting every loan as an avoided new purchase assumes an answer that the transaction record itself does not contain.

A short optional question can help establish what borrowers believe they would otherwise have done. Keep the alternatives neutral and allow uncertainty. Someone who has not considered buying the item should not be pushed into selecting that response because it makes the programme appear more beneficial. Record the answer as a reported alternative, not an independently observed event in a world that never occurred.

Keep the survey denominator visible

If forty of one hundred borrowers answer a question, the responses describe those forty participants. They do not automatically describe the remaining sixty. Report the number asked, the number answering and the basis of any percentage. A finding about respondents may be informative without being representative of everyone who uses the service. Missing answers should remain visible rather than disappearing behind a precise-looking proportion.

Avoid asking the same person to account for the full replacement value of an item on every repeat loan without considering what their answer means. If someone would have purchased a drill once, several later loans do not necessarily avoid several separate drill purchases. The relevant alternative may operate at the level of a person and a period, rather than independently for every transaction.

Distinguish visits from additional local activity

A visit to collect equipment is observable if the service records it. Whether that visit creates an additional journey to the town centre is a separate question. The borrower may already have planned to shop nearby, attend an appointment or travel past the collection point. Describing every collection as an extra town-centre trip ignores the possibility that it forms part of an existing journey.

Likewise, a nearby purchase is not established by a visit to the lending service. If local businesses or community partners want to understand linked activity, they need appropriate evidence about what visitors did, not simply a larger multiplier applied to loans. A report can discuss the opportunity for connections without presenting additional spending or business growth as an observed result.

Questions about combined trips can be modest and specific. Ask whether the visit was combined with another purpose and, if relevant, whether that other activity was already planned. Responses still reflect people's accounts rather than a controlled measurement of causation. Keeping that boundary clear allows useful local information to be shared without turning a plausible relationship into an unsupported impact claim.

Make the collection itself visible

An overall loan total can conceal substantial differences between item types. Some objects may circulate frequently, while others serve occasional but important needs. A review should connect borrowing to what was actually available. An item that spent much of the period awaiting attention cannot be compared fairly with one available throughout, simply by placing their annual loan counts side by side.

Record the periods when an item could be borrowed and the reasons it was unavailable where those reasons matter to the decision. This is an administrative description, not a substitute for appropriate equipment inspection or maintenance. It helps staff distinguish low observed borrowing from limited opportunity to borrow. The resulting question may concern availability rather than lack of interest.

Enquiries for unavailable items add another perspective. Several requests for the same object can indicate unmet interest, but repeated enquiries from one person should not be presented as several distinct potential borrowers. Keep enough context to understand the request without gathering unnecessary personal information. A concise record of the item, timing and outcome may answer the operational question more effectively than a long narrative.

Choose comparisons that answer a decision

Year-on-year comparisons become easier to interpret when the report records changes in opening hours, collection locations and catalogue size. More loans after adding a collection point may reflect increased access as well as changing interest. A shorter opening period can reduce opportunities to borrow. These conditions do not invalidate the count; they explain the environment in which it was produced.

Seasonal items also need an appropriate comparison period. A garden tool's borrowing during winter says little about its role during the growing season. Compare relevant periods before drawing a conclusion about the whole collection. If a service has only recently started, describe its early activity rather than presenting a short initial period as an established annual pattern.

The comparison should lead to a decision that the organisation can consider. It might change collection hours, clarify an item listing, improve how enquiries are recorded or investigate a frequently unavailable category. These actions address different findings. State which observation prompted the proposal so that a later review can assess whether the change helped with the identified issue.

Reconcile the records before reporting

A loan can begin in one reporting period and end in the next. Decide whether the activity measure counts issues, returns or completed borrowing cycles, and preserve that definition. Counting an issue in December and its return in January as two separate loans would describe something different from counting one borrowing arrangement. Both dates belong in the record, but their presence does not create two new loans.

Partial returns deserve similar attention. If a borrower brings back one of two items and returns the other later, the original collection remains one collection transaction while the return activity now involves two visits. Link each item to its original borrowing record. This makes it possible to count the relevant events without assuming that all items in a transaction always move together.

Renewals can also be represented differently by different systems. An extension that leaves the same object with the same person is not necessarily a fresh collection visit. Before combining exports from two locations, check how each records that extension. Otherwise, a difference in software settings may appear as a difference in service activity. Keep a brief explanation of any adjustment used to make the records comparable.

Cancelled reservations should remain distinct from completed issues. A booking may show interest in an item without showing that it left the collection. When preparing the report, check a small selection of records against their underlying events to confirm that the exported categories mean what the team expects. This is a consistency check, not a claim of an independent audit or a guarantee that every record is error-free.

If a definition changes, retain an explanation alongside the next comparison. It may be possible to restate the earlier period using the new rule, but only if the necessary underlying records exist. Where they do not, identify the break in comparability. A clear limitation is more useful than a seamless-looking trend assembled from measures that no longer describe the same events.

Publish a compact account of the evidence

A clear impact report does not need to place every working record in public. It needs enough explanation for a reader to understand what was counted, what was estimated and what remains uncertain. Keep methodological notes near the relevant claim, particularly when a headline figure depends on a multiplier or a survey response. A distant disclaimer is easy to miss.

• Name the unit: item loan, transaction, borrower or visit.
• State the reporting period and the population included.
• Separate direct records from survey answers and calculated estimates.
• Explain grouping, repeat use and missing responses where relevant.
• Show the assumptions behind any modelled result.
• Connect proposed improvements to the observations supporting them.

Check that the summary and detailed sections use the same definitions and totals. If an estimate is revised, update every place where it appears and retain a clear internal record of the change. Consistency does not prove that the method is right, but inconsistency makes a report harder to interpret and can leave readers comparing figures that were never intended to represent the same thing.

Libraries of things can describe their work convincingly without making every transaction stand for every possible benefit. Item circulation shows use of the collection. Distinct borrowers show one aspect of reach. Customer accounts help explain alternatives, while carefully defined visit records illuminate contact with the service. Together, these measures offer a fuller account than a single expansive number and a better basis for deciding what to improve next.