Matthew McKelvy · Compensation & Workforce Strategy · Pebble Beach, CA
I build the systems companies use to decide what work is worth.
Five years at Cisco advising a ~25,000-person commercial organization — compensation structure, market benchmarking, job architecture, and the offer decisions in between. When the tool didn't exist, I built it. Lately, with AI.
Cisco 2021–2026~25,000 GTM org advisedEvery BU signed my job architecture
The job. Three planes of data, one system. Grab it. Give it a spin.
One question, asked harder every year.
How should an organization structure — and pay for — its work? I started with the shape of teams, moved to the price of talent in one market, then took both at once for Cisco's commercial organization. The instruments kept upgrading: spreadsheets → models → dashboards → AI-built systems. The question never changed.
2021
Workforce & org strategy
The shape of organizations — attrition, grade distribution, org design across Sales and Finance.
2023
Regional compensation
The price of talent — benchmarking Cisco's UK pay against market for regional strategy.
2023–26
GTM compensation
Structure and price together for ~25,000 people — ranges, offers, exceptions, acquisitions.
Now
AI-era comp infrastructure
Rebuilding the machinery itself — architecture, decision tools, and workflows built with AI.
Cisco Systems Aug 2021 – Jul 2026
Compensation Analyst
Go-to-Market Organization · Los Gatos, CA
Oct 2023 – Jul 2026
Mandate Trusted comp advisor to leaders of a global commercial organization — sales and marketing, ~25,000 people — on structure, workforce planning, and resource allocation.
Led an AI-assisted job architecture refresh — hundreds of role cards structured into new families, groups, and titles — and won sign-off from comp leads in every Cisco business unit ahead of Workday implementation.
Built and shipped a modeling tool for out-of-range offer exceptions: external market data, internal range positioning, and headcount by level on one screen. Market context and internal consistency, finally in the same room.
Traced mass APJC attrition to market reference ranges that had drifted below peers; drove the full recalibration.
Built the Tableau dashboard mapping acquired Splunk roles onto Cisco's structure for post-acquisition comp decisions.
People & Communities Representative
UK & Ireland Country Consulting · Boston, MA
Feb – Oct 2023
Led market benchmarking of Cisco's UK compensation against peer data; org and workforce analyses for regional strategic planning.
People & Communities Representative
Workforce & Organizational Strategy · Boston, MA
Aug 2021 – Feb 2023
Drove a global workforce strategy program across Sales and Finance; built the business case behind Cisco's FY2022 org design playbook.
Attrition, time-in-grade, and grade-distribution analyses in Workday Adaptive Planning, presented company-wide. Chief of Staff for FLEX, Cisco's HR rotational program.
Four problems, taken apart. tap one
The attrition nobody could explain
Problem. Roles across the APJC region — GTM, operations, engineering — were bleeding people, and the losses didn't map to any team or manager.
My role. Spotted the pattern. Owned the diagnosis.
The system. Cross-referencing attrition against comp positioning showed the mechanism: market reference ranges had quietly drifted below peer benchmarks. Not culture. Pricing. I drove the full recalibration.
Value. Retention problems get expensive twice — departures, then the wrong fixes. This fixed the cause.
Signal — I look for the system underneath the symptom.
Job architecture, rebuilt with AI
Problem. Ahead of a Workday implementation, the job architecture — families, groups, titles; the skeleton every pay decision hangs on — needed a full refresh.
My role. Led the refresh, carried it through alignment.
The system. AI tooling to scrape and structure hundreds of role cards into a new architecture. Then the harder half: sign-off from compensation leads in every Cisco business unit.
Value. Architecture is load-bearing — get it wrong and every downstream decision inherits the error. Shipped aligned, on time.
Signal — AI for the volume, judgment for the taxonomy, patience for the politics.
The offer-exception model
Problem. Out-of-range offer exceptions were decided on fragmented evidence — market data here, internal positioning there, headcount somewhere else.
My role. Designed and shipped the tool myself, with AI tooling — no engineering queue.
The system. One screen: external market context, internal range consistency, headcount by level. The trade-offs visible instead of implied.
Value. Exception calls moved from assembled-by-hand to answered-on-sight.
Signal — when structure doesn't exist, I build it.
Mapping Splunk onto Cisco
Problem. An acquisition brought thousands of employees whose roles and pay structures had to reconcile with Cisco's.
My role. Built the comparison layer comp decisions ran on.
The system. A Tableau dashboard mapping acquired roles against Cisco's job structure — one shared view of how two companies actually lined up.
Value. Integration comp done badly seeds years of internal-equity problems. This made the calls inspectable.
Signal — messy org data, turned into an instrument people decide with.
Workforce planningM&A comp integrationOrg analyticsWorkday + AdaptiveTableauExcel, the deep endAI-assisted tool buildingCross-functional alignment — HR · Finance · hiring teams
Why this stack compounds now: comp and workforce data is exactly the fragmented infrastructure AI finally makes powerful — and the models still don't know the org. I've spent five years as the translation layer, and the last two building with the tools.