What Is Tail Spend? How to Define and Diagnose It
“Tail spend was never a knowledge problem — it was an arithmetic problem. AI agents change that arithmetic; what decides the outcome now is the discipline of your definitions and your data.”
| Statistic or key finding | Source |
|---|---|
| A common definitional lens puts tail spend at approximately the smallest 20% of total spend, divided among the majority of suppliers | University of Turku thesis, 2020 |
| 65% average transaction compliance, and 24% average maverick buying in services, in cited 2006 Aberdeen studies | Helsinki School of Economics dissertation, 2009 |
| 50% maverick spend reported at the UK National Health Service in a cited 2005 study | Helsinki School of Economics dissertation, 2009 |
| Five forms of maverick buying, from unintentional to ill-intentioned | Helsinki School of Economics dissertation, 2009 |
Each figure belongs to its own study population, sector, metric, and denominator; the 2005-2006 observations are preserved as historical evidence. These are definitional lenses and past measurements, not current benchmarks and not a time series — establish your baseline in your own spend analysis.
What is tail spend?
Ruutu's 2020 thesis defines tail spend as roughly the smallest 20% of spend distributed among most suppliers. Paasman's 2022 study says the definition varies by organization.
Those findings are compatible: the 80/20 curve is a useful segmentation lens, not a rule that proves where every organization's tail begins. A supplier-ranked tail, a transaction-value threshold, and spend outside category coverage can produce different totals from the same ledger. State the unit before choosing a tail-spend treatment.
Why do tail spend definitions disagree?
Definitions disagree because a useful definition matches a decision, and different organizations face different decisions. Paasman's study reports that tail spend can occur in every purchasing category and that different tactical sourcing levers may apply.
For a reproducible baseline, record the analysis period, included business units and categories, ranked unit, exclusions, and threshold. Then save the rule with the spend-analysis output. If the rule changes, recalculate the earlier period before comparing totals; otherwise the apparent movement may be a definition change rather than a purchasing change.
| Measure | Calculation | Decision it informs |
|---|---|---|
| Tail-spend share | Defined tail spend ÷ in-scope spend × 100 | Scale of the segment under the recorded definition |
| Tail-supplier share | Tail suppliers ÷ in-scope suppliers × 100 | Supplier footprint of the defined tail; interpret alongside tail-spend share |
| Tail purchase-order (PO) share | Tail purchase orders ÷ in-scope purchase orders × 100 | Operational volume created by the tail |
| Average tail PO value | Tail spend ÷ tail purchase orders | Whether workflow effort is proportionate to order value |
| One-time supplier rate | Tail suppliers with one purchase order ÷ tail suppliers × 100 | Possible consolidation or data-quality questions |
| Applicable-contract leakage | Non-compliant tail spend where an established contract or required process applies ÷ tail spend subject to that contract or process × 100 | Where established buying channels are bypassed; diagnose the reason before labeling it maverick buying |
| Repeat-demand share | Recurring tail purchase orders ÷ tail purchase orders × 100 | Catalog or consolidation candidates |
| Competition coverage | Eligible tail events with more than one qualified bid ÷ eligible tail events × 100 | Whether eligible demand is reaching multiple qualified suppliers |
Use one analysis period, scope, currency, tail definition, and data-cleaning rule throughout. Define a normalized supplier, recurring demand, an eligible event, and a qualified bid before calculating their shares. If a denominator is zero, report the measure as not applicable rather than zero.
What does the older research still explain?
The older research in this review remains useful because it explains behavior. A 2009 peer-reviewed paper used a systematic literature review and interviews to identify forms of maverick buying. A 2011 peer-reviewed study tested contributing factors with survey data from Finnish governmental procurement.
These studies help a buyer distinguish reasons for non-compliance; their age does not erase that conceptual value. Their settings and methods also mean they do not establish a current prevalence rate for a multi-site industrial company.
A 2009 dissertation also preserves two older, context-specific observations: cited Aberdeen studies reported 65% average transaction compliance (2006) with maverick buying in services averaging 24%, and a cited 2005 UK National Health Service study reported 50% maverick spend. The populations, sectors, metrics, and denominators differ, so the figures cannot be combined into a time series. Keep them as a record of what procurement teams were measuring, then establish a current internal baseline in your own spend analysis.
What does newer research add?
Newer evidence adds questions about classification and process design rather than a replacement definition. Guida, Caniato, and Moretto's 2025 study uses a provider-focused case-study approach to examine AI-based spend classification and explicitly notes a gap around human oversight. APQC's 2025 public summary links consistent procurement processes with preventing maverick buying and says its findings draw on procurement benchmarking research.
Neither source supplies a universal tail-spend rate. Together they support a practical position: use tooling to organize the data, but keep the definition, exceptions, and category decisions visible in the spend-analysis workflow.
Why does tail spend escape procurement?
Small purchases create a process-fit problem. Ruutu notes that a compliant process can be excessively heavy for tail purchases. APQC's current summary says consistent procurement processes can prevent buying outside organizational standards.
Those points are not opposites: standardize the decision rights and data capture, while matching the effort to the purchase. In urgent, locally bought categories such as maintenance, repair, and operations (MRO) and facilities, ask where the intended workflow creates avoidable delay and which controls can be simplified without losing visibility.
Is tail spend the same as maverick buying?
No — the two overlap but describe different problems. The 2011 study defines maverick buying as non-compliant, off-contract purchasing where an established process and pre-negotiated contracts already exist. The 2009 paper shows that the underlying forms can range from unintentional behavior to deliberate sabotage. A compliant low-value order can still belong to the tail, while an off-contract order can reflect lack of awareness, a process mismatch, or a deliberate choice — so measure the buying channel and the reason before assigning a remedy in the tail-spend operating plan.
What should procurement leaders do about tail spend?
- Write the definition at the top of the spend analysis: period, scope, ranked unit, exclusions, and threshold. If those fields differ, do not compare two tail estimates.
- Break the total into categories and buying channels. In MRO and facilities, separate compliant small orders from purchases outside an intended contract or workflow.
- Count suppliers, purchase orders, repeat purchases, one-time suppliers, and orders with more than one qualified bid. These observations help distinguish fragmentation, channel compliance, process burden, and a competition gap.
- Attach one treatment hypothesis to each category: compete, consolidate, catalog, simplify, or control. Use a tail-spend operating plan to record the owner, baseline, decision rule, and review date.
Where to start: one category, one decision rule
Choose one category only after the diagnosis. If specifications are comparable and the current shortlist is narrow, a bounded competitive event may be a useful test. If purchases repeat and requirements are standard, consolidation or a catalog may remove more friction.
If urgency dominates, a controlled channel may fit better. Record the supplier set, transaction count, buying channel, internal effort, and commercial baseline before the change, then apply the same measures after the pilot. Expand the tail-spend program only when that category evidence supports the decision.
How do AI agents change tail spend management?
Frequently asked questions
Does maverick buying cause tail spend, or result from it?
Both directions appear in the research: the University of Turku thesis recognizes maverick buying as a driver of spend fragmentation but also a consequence of it and of uncontrolled tail spend. A remedy that fits one direction can worsen the other, so diagnose the direction first.
Is maverick buying always intentional?
No. A 2009 peer-reviewed paper identifies forms of maverick buying ranging from unintentional behavior to straightforward sabotage, so the effective remedy depends on the reason for non-compliance, not only on detecting it.
Do rewards and sanctions reduce maverick buying?
A 2011 survey study of Finnish governmental procurement found that training can help prevent maverick buying, while reward and sanction systems did not appear helpful for the three forms it tested. Diagnose the form of non-compliance before choosing a control.
Why does maverick buying raise costs?
Sources
- Defining and categorizing tail spend for a better management thereof — T. Paasman, University of Twente, 2022. Foundational evidence (masters thesis): Why a tail-spend definition must be organization-specific, and why treatment should be chosen at category level.
- How to manage the long tail of procurement — A. Ruutu, University of Turku, 2020. Foundational evidence (masters thesis): The supplier-distribution definition and the process-weight mechanism that can contribute to maverick buying.
- Non-compliant work behaviour in purchasing: An exploration of reasons behind maverick buying — K. Karjalainen, K. Kemppainen, and E. M. van Raaij, Journal of Business Ethics, 2009. Foundational evidence (peer reviewed journal): The range of maverick-buying forms and the fact that different reasons require different diagnoses.
- An empirical test of contributing factors to different forms of maverick buying — K. Karjalainen and E. M. van Raaij, Journal of Purchasing and Supply Management, 2011. Foundational evidence (peer reviewed journal): A precise definition of maverick buying and evidence that remedies depend on the reason for non-compliance.
- AI meets spend classification: A new frontier in information processing — M. Guida, F. Caniato, and A. Moretto, Journal of Purchasing and Supply Management, 2025. Current empirical evidence (peer reviewed journal): The current information-processing challenge in spend classification and the continuing need for human oversight.
- Apply Consistent Procurement Processes to Reduce Maverick Buying — APQC, 2025. Contextual evidence (benchmarking research): A current process-standardization perspective, without importing member-only or undisclosed numeric benchmarks.
- Challenges of purchasing centralization: Empirical evidence from public procurement — K. Karjalainen, Helsinki School of Economics, 2009. Historical evidence (doctoral dissertation): Two explicitly historical compliance observations preserved as context, not as a current rate or time-series baseline.