Selected work.
Real engagements, real numbers, real lessons.
Scaling BI through 3.4× growth.
250 → 850 restaurants over 8 years. 15 TB migrated to Snowflake. Tableau company-wide in 120 days.
Read →Rebuilding loss-mitigation analytics.
80%+ data-access reduction. Credit Data Repository for the entire loss-mitigation organization. Self-service Analytics Portal in active use.
Read →Finding $5M with the right question.
$5M in unneeded orders identified — and avoided — by analyzing on-hand, on-order, and historical sales data together for the first time.
Read →From 10% to 90% on-time reporting.
10% → 90% on-time delivery in 18 months. Weekly warehouse processing: 12 hours → 4 hours.
Read →Fast-growing QSR chain — Scaling BI through 3.4× Growth
Situation
In 2016, the chain was 250 restaurants and growing fast. The BI function was working, but the architecture, the team, and the delivery cadence were all sized for a company that no longer existed. The board’s growth plan would 3× the company in eight years — and the analytics function had to keep up without becoming the bottleneck.
What we did
- Stabilized the data warehouse and reporting environment. Established SLAs, fixed top quality issues.
- Migrated the entire 15 TB EDW from on-prem SQL Server to Snowflake on AWS via SnapLogic ELTL.
- Replaced Sisense with Tableau company-wide in 120 days. Built role-aware data security with SSO and automation.
- Founded an IT Quality Assurance function from scratch — hiring, processes, operationalization.
- Built franchise-partner data pipelines for national QSR franchise partners as ‘lights-out’ systems.
- Created a single view of customer across order channels (mobile, dine-in, drive-thru).
Outcome
- Scaled from 250 to 850 restaurants over 8 years.
- Information delivered to business leaders 2 hours earlier than prior SLAs.
- Annual contractor spend reduced by $200K while supporting 20% annual growth.
- Tableau deployed company-wide in 120 days.
- Single customer view across all order channels.
Lessons
- Don’t try to migrate the platform and rebuild the team at the same time. Sequence them.
- Tableau is the easy decision; the hard decision is the governance model around it.
- A QA function pays for itself within a year. Build it before you think you need it.
Mortgage finance enterprise — Rebuilding Loss-Mitigation Analytics
Situation
The enterprise’s loss-mitigation organization needed analytics it could trust during the years following the housing crisis. The team existed; the processes did not. Data access was slow, contractor spend was high, and self-service was nonexistent.
What we did
- Realigned a 25-person team. Tightened priorities, refocused capacity, reduced contractor reliance.
- Optimized Oracle and Netezza platforms; reduced data-access times by over 80%.
- Led development of the Credit Data Repository — a single source for the whole loss-mitigation organization.
- Founded the Data Visualization team. Built a SharePoint-based Analytics Portal for self-service.
- Led the STAR (Servicer Total Achievement and Rewards) analytics program.
Outcome
- Contractor spend reduced 40% through team realignment and productivity improvement.
- Data-access times reduced by 80%+.
- Credit Data Repository serving the entire loss-mitigation organization.
- Self-service Analytics Portal in active use across the function.
Lessons
- Self-service in a regulated environment requires governance first — there is no shortcut.
- The fastest way to reduce contractor spend is to give your FTEs better tools, not more pressure.
Supermarket distributor — Finding $5M with the Right Question
Situation
The client was a major supermarket distributor with effectively no enterprise data warehousing capability. Buyers, distribution centers, and merchandisers all worked in their own data. The company assumed it had no opportunity to consolidate — nobody had asked the question that would have proved otherwise.
What we did
- Built an enterprise data warehouse providing a 360° view of suppliers, distribution centers, and customers.
- Centralized all supply-chain data — inbound and outbound orders, inventory, customer fulfillment.
- Connected on-hand, on-order, and historical sales data into a single analyzable model.
Outcome
Lessons
- The biggest savings live in data nobody connected before.
- Distribution analytics is supply-chain analytics is operations analytics. The org chart should not pretend otherwise.
Digital marketing agency — From 10% to 90% On-Time Reporting
Situation
Inherited a poorly-performing analytics operation inside a global holding-company agency. Weekly reporting hit deadline 10% of the time. Production data warehouse processing took 12 hours, eating into the window for QA. Individual contributors were good; they were not operating as a team.
What we did
- Rebuilt delivery cadence and accountability structures.
- Restructured ETL and warehouse load processes — 12-hour cycle to 4 hours.
- Coached a group of individual contributors into a collaborative delivery team.
Outcome
Lessons
- Most missed deadlines are not technology problems. They are operating-model problems.
- Eliminate ambiguity about what ‘done’ means — then performance follows.
Want to compare notes on a similar situation?
Most of the work we do is a variation on these four patterns. The fastest way to know if we’re a fit is a 60-minute conversation.