Analytics & AI consulting · San Francisco

Data that changes decisions, not just dashboards.

I'm James Oliver. I spent 15 years building and running analytics teams at Rakuten Rewards, The RealReal, and Comcast. Now I help companies measure what's working, cut what isn't, and put AI to work on their data.

Book a 30-minute call See services
Experiment readoutExample

New offer page vs. current page

MetricControlTest
Conversion rate3.10%3.29%
Revenue per visitor$1.84$1.93
Sample size48,21048,377

Conversion lift +6.1% (95% CI +2.2% to +10.0%)

Ship it. The lift is real and worth about $430K a year at current traffic.

Millionsin annual revenue recovered with partner incrementality testing
+24%more bookings in two months for a small-business client
12person global analytics & engineering team built from the ground up
15 yearsleading analytics for retail, marketplaces, ad tech, and media

Clear answers, at any size.

From a one-person business to a major cash back platform, the approach is the same: measure what's working, then do more of it.

Client · Personal chef

Which marketing actually brings in clients?

From Scratch Personal Chef wanted more clients and wasn't sure where to start with analytics. I set up tracking across the website and marketing channels, showed which ones brought in bookings, and focused the effort there.

+24%bookings in two months
+87%website visits
"James was super straightforward and helpful. He showed me which marketing channels I should pursue and why."From Scratch Personal Chef logoElizabeth Long, From Scratch Personal Chef

Rakuten · Department store

Does cash back actually drive new revenue?

For a high-end department store, my team designed and analyzed a randomized test that gave shoppers 0%, 2%, 5%, or 10% cash back, to measure the sales that wouldn't have happened without it.

13%incremental revenue lift
+61%sales per buyer
+64%average order value
Read the case study at Rakuten →

Rakuten · Fashion brand

Finding new customers with machine learning

For a high-end fashion brand, my team analyzed a personalized 20% cash back offer sent only to new audiences picked by a machine learning model, with existing customers left out.

2×return on ad spend
111%sales lift on media days
3.5×email click rate vs. standard
Read the case study at Rakuten →

Public sector · ConfidentialAnalytics work for the State of California. The project details are confidential, and I keep them that way.

Rakuten figures are from Rakuten Rewards' published case studies, from my team's analyses there. Rakuten is a former employer, not a client.

Two ways to work with me

No data team? I'll be yours. Already have one? I'll make it stronger.

Small businesses

I run your analytics for you

Not sure what your website traffic or social accounts are telling you? I set up the tracking, read the numbers every month, and tell you in plain English what's working, what isn't, and what to try next.

Larger organizations

A consultant inside your organization

I work alongside your leaders and your data team, turning analytics from a place that answers questions into a partner in every major decision, with clear numbers on how each choice is likely to play out.

Industries

Retailers, brands, and the platforms between them.

I've spent 15 years inside marketplaces, media companies, and ad platforms. I'm especially useful to established companies that have plenty of data and want clearer answers from it.

Consumer brandsCPG brandsProving which retail media, promotions, and partner programs actually drive sales.
RetailBig-box and specialty retailersOmnichannel measurement, loyalty, and self-serve reporting for large merchant teams.
Retail mediaRetail media networksIncrementality testing that shows advertisers the real lift behind their spend.
LoyaltyCash back, loyalty, and affiliatePartner value, offer testing, and personalization, from nine years at Rakuten Rewards.
MarketplacesMarketplaces and luxury resaleConversion, segmentation, and experimentation, from The RealReal and Rakuten.
MediaCable, telecom, and mediaProduct usage and customer insights, from Comcast.
AdvertisingAd tech and publishersKPI frameworks and network performance, from Technorati.
Established companiesLegacy enterprisesModernizing reporting and data costs without a big rebuild.

Services

Six things I do well.

Each one comes straight from work I've done inside large companies, sized for teams that don't have a head of analytics yet, or need a second opinion from one.

Experimentation

Know what actually works

A/B tests and incrementality studies designed properly, so you stop paying for results you'd have gotten anyway.

  • Test design, sizing, and readouts
  • Partner and channel incrementality
  • Experimentation programs your team can run
Analytics strategy

A fractional head of analytics

Senior leadership for your data team, part-time. I set the KPIs, the roadmap, and the standards, and coach the people doing the work.

  • KPI frameworks and metric definitions
  • Team structure and hiring
  • Exec-ready reporting
Data platform

Faster data, smaller bills

Clean models, one source of truth, and self-serve dashboards people trust, with Snowflake spend under control.

  • Snowflake cost reviews
  • Data modeling and semantic layers
  • Looker and Tableau rollouts
AI enablement

Put AI to work on your data

Practical AI for analytics teams: the right use cases, prompts that hold up, and guardrails so the numbers stay right.

  • AI use-case roadmaps
  • Prompt libraries and team training
  • Measuring what AI features actually deliver
SQL

SQL that does the heavy lifting

Most business questions can be answered with well-written SQL on the data you already have. I write it, review it, and make it faster and cheaper to run.

  • Query reviews and rewrites (1,000 lines down to 300)
  • Faster, cheaper Snowflake queries
  • Clean data models your team can own
Team leadership

One team from many backgrounds

I've built and led analytics teams across countries, cultures, and disciplines, and I'm good at getting engineers, analysts, and business leaders with very different perspectives to agree on one set of numbers.

  • Global teams across the US, Canada, and Japan
  • Hiring, coaching, and team structure
  • Shared metric definitions across markets

Our approach

Analytics, made simple.

You don't need a data science team, a fancy model, or a big budget to get answers you can act on. Clean data, well-written SQL, and a properly designed test get companies most of the way there. We start simple, and only add complexity when it clearly pays for itself.

Simple doesn't mean small. In most companies, analytics is a help desk: someone asks a question, the team sends back a number, and that's the end of it. I move the team from the help desk to the leadership table, where it works as an internal consultancy: part of every major decision, showing with clear data how each choice is likely to affect results before it's made.

How we work

Start small. Keep going if it's working.

Most clients start with a diagnostic, then choose a project or ongoing support. You always know the price before we start.

2 weeks · from $6,000

Diagnostic

I review your data, dashboards, tests, and costs, then hand you a written plan with the three changes worth the most money.

4–12 weeks · fixed price, quoted up front

Project

One clear outcome, like an incrementality study, a Snowflake cost cut, a KPI framework, or an AI pilot.

Monthly · from $8,000 a month

Fractional leadership

A set number of days each month as your head of analytics: strategy, reviews, hiring, and exec reporting.

James Oliver, smiling, in a navy blazer and light blue shirt

Experience

Fifteen years inside the data.

Most recently Senior Director of B2B Analytics at Rakuten Rewards, where I built the analytics function into an in-house consultancy for the C-suite.

MBA, University of Utah BA in English and German, Grinnell College First place, Adobe Web Analytics Competition (2011) Case study author, Opportunities and Challenges of Workplace Diversity (Pearson, 2012)
  • SQL
  • Snowflake
  • Python
  • Looker
  • Tableau
  • Redshift
  • AWS
  • Adobe Analytics
  • Google Analytics
  1. 2017 – 2026
    Rakuten RewardsSenior Director, B2B Analytics. Built a 12-person global team; recovered millions in revenue through incrementality testing; cut data warehouse costs; reporting for an AI-driven cashback platform.
  2. 2016 – 2017
    The RealRealBusiness Intelligence Manager. Led the company-wide Looker rollout, A/B testing, and customer segmentation for a luxury marketplace.
  3. 2014 – 2015
    Technorati MediaAnalytics Manager. Built the KPI framework and Tableau reporting suite for an ad network.
  4. 2011 – 2014
    ComcastSenior BI Analyst. Product-usage insights that helped cut technical support calls by 40%.
Baxter, a tabby cat, sitting in a desk chair behind an open laptop as if in a meeting

Why "Baxter & James"

Baxter is the cat. He's the cool one.

Baxter supervises every engagement from the back of the couch. He doesn't do SQL, but he's an excellent judge of whether a meeting could have been an email. I do everything else.

Contact

Let's look at your data.

Tell me what you're trying to do. I'll tell you how I can help and what it will take.

Book a timePick a 30-minute slot →
Emailjames.somers.oliver@gmail.com
LinkedInlinkedin.com/in/jamessomersoliver
Phone(978) 675-1484
Based inSan Francisco · working with teams anywhere