Skip to presentation
Dunder Mifflin DirectQ2 2026 · Fictional research
What should we do?

Let customers shop before asking them to qualify.

Prove assortment, price, and delivery value first. Collect only what the selected purchasing path requires.

Loss accountingSignup starters who do not order
Where does failure accumulate?
6,088
signup starters do not submit a first order.

Nearly nine in ten failures occur before product value.

The aggregate problem is primarily pre-catalog. A distinct confidence problem remains after catalog access.

5,460 · 89.69%Never reach the catalog: access, profile, setup, approval, or catalog-entry loss.
628Reach catalog; no order
Starter → account −1,216Verification −504Profile −1,480Setup −1,565Approval −560Catalog entry −135First order −628
Activation journey12 segments · 12,930 visitors
Exact funnel diagnosis

Purchasing setup is the largest post-signup break.

34.55%

1,565 customers are lost between business-profile and purchasing-setup completion.

Site visitors12,930100%
Signup started7,73059.78%
Account created6,51484.27% prior
Email verified6,01092.26% prior
Business profile4,53075.37% prior
Purchasing setup2,96565.45% prior
Approved2,40581.11% prior
Catalog viewed2,27094.39% prior
First order1,64272.33% prior
Diagnostic splitDo not collapse these into one problem
What changes after catalog access?

Administrative gating dominates—but it is not the whole story.

5,460

Value is withheld

Failed starters never reach the catalog. Interviews consistently describe wanting products, prices, availability, substitutions, or delivery before tax, credit, and approval work.

628

Confidence is insufficient

Catalog viewers do not order. Price correctness, paper specifications, samples, substitutions, checkout policy, and delivery promises become decisive.

55.3% creative-agency drop39.6% print-shop drop20.0% contract-customer drop

The first problem supports ungating. The second requires segment-appropriate decision support and transaction integrity.

Cause modelObserved + interpreted
Why is it breaking?

Five mechanisms reinforce abandonment and distrust.

The funnel locates loss. Interviews and tickets suggest—but do not prove—the mechanisms.

10 interviews across 10 customer types independently asked for value before setup.
Sequence

Supplier onboarding arrives before product, price, and delivery confidence.

Distributed work

One user is expected to know finance, tax, locations, roles, and policy—with weak save or delegation.

Price trust

Mismatch ranges from 7% to 24%; legal entity, currency, freight, tax, and volume source are unclear.

Product risk

Specifications, samples, historical SKUs, and substitutes are inadequate for creative and print decisions.

Transaction integrity

PO validation, tax changes, card/order reconciliation, lost state, and permission failures create financial risk.

Product modelRecommended operating sequence
The strategic change

One storefront. Three purchasing paths.

Do not make customers choose a buying system before they can judge the offer. All three paths preserve the same cart.

Browse, compare, and build a cart immediatelyPublic or estimated price · availability · delivery · samples
01 · Lowest friction

Shop now

Lightweight account and card purchase. Defer credit and organizational setup.

02 · Existing relationship

Link an account

Match identity, legal entity, contract, currency, tax, freight, and price source.

03 · Organizational controls

Configure purchasing

Complete tax, credit, PO, locations, roles, and approvals asynchronously.

Customer systemsComparison, not hidden interaction
Who experiences it differently?

The primary barrier changes with the buying system.

Customer systemSignup → orderPrimary lossBarrierResponse
All non-contract18.21%Setup · 39.40%Value arrives after administrative commitment.Ungate catalog and persistent cart.
Complex institutions19.46%Setup · 43.37%Tax, PO, locations, roles, and approvals.Templates, delegation, asynchronous setup.
Creative & print19.84%Catalog · 47.37%Specifications, samples, SKU equivalence.Guided + precision product tools.
Existing contract43.96%Catalog · 20.0%Entity and contract-price trust.Explainable account matching.

Broad groups are directional. The appendix retains all 12 supplied segment records.

Named evidenceAll identities fictional
Four failure modes, four customer systems

The same workflow creates different kinds of risk.

Shop now · INT-001

Aisha Bennett

Maple & Main Hardware
“Legal entity, tax ID, and payment terms before it showed me a single price.”
Local-business profile completion loses 45.5%.
Configure · INT-008

Omar Moore

Greenway Health Network
“Let me build a sample cart and estimate delivery.”
One payer, many destinations; setup blocks evaluation.
Product risk · INT-026

Theo Banerjee

Northline Brand Works
“Search recognizes A4 but not the internal SKU.”
Creative buyers lose 55.3% after catalog view.
Link · INT-012

Rohan Roy

Sterling Office Products
“The portal treated us like a new company.”
Existing relationship is not reliably recognized.
Support topology120 selected fictional tickets
Evidence beyond interviews

Support reveals broken state, identity, control, and recovery.

31.67% high or urgent72.50% confused, angry, or frustrated1 positive ticketCounts are quota-balanced—not prevalence.
CounterevidenceWhat keeps the conclusion honest
94.39%approved → catalog
72.33%catalog → first order
60.0%contract 30-day reorder
The core appears capable

Value emerges once the account is connected correctly.

“Once support linked our contract, reordering was faster than emailing the branch.”

Rina Scott · Maple & Main Hardware · INT-061 · fictional

Contradictory evidence

Contract customers convert best despite the highest mismatch rate, 24%. Established demand may buffer defects. Approval speed also does not fully explain activation: creative agencies wait 31 hours and activate at 14.6%; enterprise waits 83 hours and activates at 26.1%.

Validation portfolioTest mechanisms, not opinions
What should we test next?

Lead with ungating; test trust and product confidence in parallel.

01 · Primary

Ungated catalog + cart

Email-only browsing, pricing, availability, delivery estimate, and persistent cart.

Baseline
10.68% visitor → order
Primary
First order
Guardrail
Price leakage, tax, margin
02 · Trust

Existing-account matching

Invoice/customer-number match with entity and landed-price explanation.

Baseline
24% mismatch
Primary
Time to exact price
Guardrail
False matches, exposure
03 · Confidence

Guided finder + SKU tools

Practical comparison for creatives; historical SKU equivalence for print.

Baseline
44.68% creative catalog → order
Primary
Add-to-cart, order
Guardrail
Returns, bad substitutes
Delivery sequenceProduct + operations + instrumentation
How should we move?

Sequence the work around learning and transaction safety.

NOW · 0–30 days

Stabilize and instrument

  • Add cart, checkout, payment, validation, and confirmation events.
  • Rescue stalled high-intent accounts.
  • Create one matching and exemption queue with ownership.
  • Fix card-without-order, tax-change, state-loss, and permission defects.
NEXT · 31–60 days

Run the primary test

  • Launch ungated browse-and-cart experiment.
  • Preserve carts through signup and approval.
  • Show price source and delivery estimate clearly.
  • Measure first order and guardrails.
THEN · 61–90 days

Pilot the paths

  • Card-first for eligible buyers.
  • Invoice/customer-number account matching.
  • Institutional setup templates and delegation.
  • Guided finder and historical SKU lookup.
Leadership implicationCore deck close
The decision

Move proof of value ahead of administrative work.

ApproveUngated catalog-and-cart experiment for non-contract visitors.
InstrumentThe missing cart, checkout, payment, and validation events.
PrepareCard-first, account-linking, and organizational setup paths.

The opportunity is not to replace the ordering core. It is to let more customers experience it.

Appendix · Segment dataQ2 2026
All supplied segments

Activation, primary loss, approval, mismatch, and support.

SegmentSignup → orderLargest lossApprovalMismatchSupport / 100
Local small business11.1%Profile · 45.5%44 h12%23.4
Creative agency14.6%First order · 55.3%31 h8%18.8
Independent print shop26.9%First order · 39.6%26 h10%16.1
Mid-market office24.4%Setup · 35.1%58 h17%21.7
Enterprise procurement26.1%Setup · 43.9%83 h21%28.5
K–12 education13.5%Setup · 51.4%91 h9%30.2
Higher education18.0%Setup · 48.4%88 h14%27.6
Healthcare network26.0%Setup · 38.2%69 h13%24.0
Government agency12.3%Setup · 56.8%112 h11%34.1
Hospitality group19.0%Setup · 35.5%61 h15%25.8
Nonprofit12.0%Setup · 50.0%77 h7%29.7
Existing contract44.0%First order · 20.0%19 h24%12.4
Appendix · Secondary findingsImportant, not the primary decision
What else matters?

Activation is primary; retention and transaction quality remain exposed.

Repeat ordering

51.46% reorder within 30 days. Shared lists, copy order, CSV recovery, and substitution consent are incomplete.

Delivery confidence

Stock and promised dates change after payment. Post-purchase contains 14 high or urgent tickets out of 20 selected.

Accessibility and mobile

Keyboard-inaccessible filters, repeated screen-reader prices, covered cart controls, and unreadable warehouse timelines.

Returns and substitutions

Wrong brightness, damage, lot variation, missing credit memos, and unauthorized substitutes rationalize pre-order concern.

Organizational controls

Overbroad location access and bypassed thresholds are security and internal-control risks—not mere friction.

Quick order

The strongest positive pattern: experts succeed when the SKU is known. Preserve it while adding discovery and SKU mapping.

Appendix · LimitsDirectional, not causal
Method and unresolved questions

The direction is credible; the mechanism and effect size require testing.

Research base: Q2 funnel across 12 segments and 12,930 visitors; 62 interview excerpts; 120 selected tickets. All identities and records are fictional workshop material.

The funnel cannot distinguish a hard catalog gate from practical abandonment.
There are no add-to-cart, checkout-start, payment-attempt, order-value, or abandonment-reason events.
The ticket sample is quota-balanced and cannot estimate natural prevalence or resolution rate.
The denominator for contract-price mismatch is not supplied.
“First order” may mean first portal order rather than first-ever Dunder Mifflin order.
Portal abandonment may move orders to branch representatives rather than competitors.
Legal requirements must be separated from inherited internal process.
Thirty-day reorder may not match every category or segment cadence.