Increasing operational capacity without increasing headcount
E-commerce / Liquidation auctions · Redesigning case ownership to increase operational capacity · via Hugo
The team owned only a minority of case volume, so demand kept escalating outward and capacity looked like a headcount problem.
Most escalations were repeat product questions and unclear ownership, not genuinely complex cases — 12% of volume was product-question swarm.
Redesigned case ownership, standardised process, rebuilt the knowledge base, targeted overtime deliberately and wrote the back-office operating playbook.
Case volume owned rose from 37% to 72% with fewer than 3 open items at end of shift, and product-question swarm fell from 12% to 1%.
Overview. The client runs online liquidation auctions — returned, overstock and freight-claim merchandise sold through HiBid, with operations run out of Auction Flex. Hugo provided customer support (Salesforce cases and calls) and, as the engagement matured, back-office admin operations: bidder approvals, payment processing, auction creation and lot uploads.
I managed a team of 5, building the QA framework and admin operations playbook from scratch, owning workforce management, and managing a client relationship that was by design exacting and often difficult — the client's whole model runs on strict enforcement of its own Terms & Conditions.
- Managed 5 agents across customer support and, later, back-office auction administration.
- Designed the client's QA scorecard from scratch — 100 points across four dimensions — calibrated to the client's specific policies.
- Documented the full admin operations playbook so back-office work ran as repeatable process, not tribal knowledge.
- Owned scheduling and approved overtime against specific, named backlog goals rather than open-ended extra hours.
- Managed a demanding client relationship, translating a policy-first culture into rules agents could apply consistently.
The account asked the team to stand up two different functions at once: real-time customer support and precise, rules-heavy back-office auction administration, for a client with an unusually strict operating culture.
The client's own rubric instructs agents not to apologise by default “especially when a customer is in the wrong”; policy did not allow combining orders, cancelling won bids, using outside couriers, or refunds outside narrow reason codes. Agents were routinely absorbing customer frustration with policies they had no discretion to bend — while the client relationship itself was exacting, with little margin for error.
| Dimension | What it measures |
|---|---|
| Empathy (25) | Polite, respectful, calm tone — calibrated to the client's philosophy of apologising only when genuinely warranted |
| Average Handling Time (25) | Email responded to within 10 hours; calls not abruptly ended |
| Accuracy & Resolution (25) | Correct, complete resolution; templates current; escalate only after attempting resolution |
| Written Communication (25) | Clear, professional, brand-aligned language; proper formatting; correct call openers |
Calibrated to this client, not a generic rubric
The Empathy dimension is the clearest example: it explicitly scores agents down for over-apologising, which runs counter to how most CX QA frameworks are built, because the client's brand voice prioritised enforcing its stated terms over customer accommodation.
| Process | What it covers |
|---|---|
| Auction creation & upload | Building a new auction in Auction Flex from a prior closed auction's settings, then uploading to HiBid |
| Bidder approval / decline | Approving Canadian bidders with acceptable credit scores; declining others with a documented reason |
| Payment processing | Charging cards post-auction, retrying up to 4 days, then permanently blocking non-payers |
| E-transfer logging | Matching manual payments to invoices and updating fulfilment status |
| Returns & refunds | Return orders for fulfilled shipments only, with reason codes and status updates |
| Price auditing | Cross-checking flagged lot prices against Amazon CA, Best Buy and Wayfair |
| Daily lot / bid uploads | Cleaning cataloged data in Modern CSV, importing lots and images, flagging cataloging errors |
Making back-office work scalable
Documenting the playbook is what let the team take on the account's Admin Process Handling Phase as a genuinely new line of work rather than an ad hoc favour — down to granular rules like exact bid-to-reserve spacing ($1 below reserve under $30, $2.50 below reserve $30–$100, $10 below reserve $100–$1,500) and rotating bidder accounts to avoid the appearance of shill bidding.
Managing workforce and backlog with targeted overtime
I tracked daily shift metrics — calls, closed cases, new swarms, cases awaiting updates, escalations, bid permissions, auction uploads — so backlog was visible in real time rather than discovered after the fact. Overtime wasn't blanket-approved: it was logged against a specific, named purpose (most often “reduce email count on service queue”) with hours and outcomes tracked, so extra hours translated into a measurable dent in the backlog.
Managing a demanding client relationship
The client's strict policy environment created recurring friction — customers pushing back on stated policies, time-bound shipping cases stuck with an unresponsive shipping department, and at one point customers hearing incorrect warehouse hours from the client's own automated phone system.
Rather than let agents absorb that friction improvised, I built structure around it: documented, reason-coded responses for recurring edge cases, a 'swarm' process to jump on time-bound cases as a team, and direct escalation of client-side problems instead of leaving agents to repeatedly explain a client mistake to frustrated customers. A meaningful part of the job was absorbing that difficulty so it didn't land on the agents or the customers.
| Metric | Result |
|---|---|
| Share of case volume owned | Grew from 37% to a sustained 69–77%; 72% in the final two-week period |
| Avg inbox status at end of shift | Held under 3 open items across every week measured |
| 'Product Questions' swarm share | Fell from 12% to 1% as product knowledge improved |
| Bi-weekly volume snapshot | 1,019 cases, 590 calls, 238 escalations handled |
| QA scorecard | 100 points across 4 dimensions, calibrated to client policy |
- Share of case volume owned37% → 72%
- 'Product Questions' swarm share12% → 1%
- End-of-shift open items<3 every week
- QA framework and admin playbookBoth outlasted the role
- Took the account from handling roughly a third of case volume to closing the clear majority of it, without backlog creeping up.
- Built the QA rubric and admin playbook as reusable infrastructure that let the account expand into back-office work.
- Reduced ambiguity through process and documentation instead of leaving agents to navigate a hard client case by case.
More case studies
Full index- 01US digital pharmacy clientDiagnosing the root causes of customer and compliance failures, separating agent performance from systemic issues, and building the governance to improve both safely at scale93% CSAT vs 90% target
- 02EdTech learning marketplaceReducing recurring customer dissatisfaction through frontline capability84% drop in issue-recognition DSATs
- 03Conversational commerce SaaS platformDesigning and scaling a new customer-support channel90% onboarding pass rate vs 60% target