You bought the analytics tool. Now get the decisions.
Liftpoint Analytics is a fractional Head of Product Analytics & ML: a senior analytics lead two days a week, scoped as fixed packages with a named decision at the end of each one.
The lead offers below suit product and marketplace companies. Any business with data and a decision to make is welcome: the work is scoped and priced to fit.
Three things the role covers
Each one has a number behind it from prior work at a publicly traded two-sided marketplace for high-value physical goods.
-
Experimentation at scale
Design the test, read it honestly, and get the decision signed off by the executive who asked for it.
30+ A/B tests run. CUPED variance reduction cut test durations by about 25%. A listing-page redesign shipped to 100% of users after the exit-rate scare turned out to be a funnel shift.
-
Instrumentation ownership
Own the tracking plan, the event taxonomy and the vendor relationship, so the data is trustworthy before anyone reads it.
GA4 tag-manager tagging was losing an estimated 15–20% of data. Replaced with a digital-experience platform in about 6 months, adopted by 9+ teams. Issue resolution went from about 3 days to 1.
-
ML and AI automation that pay for themselves
One decision, one model or pipeline, a dollar outcome someone can audit. AI is used where it saves analyst time: classifying text, automating reports, watching metrics.
Break-even modeling on a delivery pilot moved programmed margin from 5% to 15% and take rate from about 0.4% to about 3% of eligible listings in about 14 months. Gradient-boosting pricing work tied to seven figures in annual savings. AI classification of about 300,000 customer conversations into about 100 reasons, and a review-sentiment dashboard used by customer success and brand marketing.
If any of these sound familiar
These are the sentences product leaders say in the first ten minutes of a call.
- “We pay for Fullstory and I'm not sure anyone opens it.”
- “Every A/B test ends in an argument about whether it worked.”
- “The CEO asked for a number and three dashboards gave three answers.”
- “Our analyst is great at SQL but can't present to the exec team.”
- “We can't justify a $180K analytics hire yet, but we're flying blind.”
- “We know pricing is leaving money on the table. We don't know how much.”
Four ways to work together
Fixed fees or monthly retainers, not open-ended hourly billing. Start with the Diagnostic unless you already know which decision you need.
-
Offer 1 · Start here
Analytics Diagnostic
$6,500 fixed
2 weeks. Credited against A or B if signed within 30 days.
Two weeks of read-only access and interviews. Where the data is, what tracking is broken, which dashboards matter, how decisions actually get made, and the top three decisions a model could improve.
- Up to 6 stakeholder interviews
- Tracking health check on up to 20 core events
- Maturity score, 90-day roadmap, written report and a 60-minute readout
-
Offer A
Fractional Head of Product Analytics
from $9,000 per month
Two days a week. 3-month minimum, then month to month. $9,000–12,000.
A senior analytics lead who owns the roadmap, designs and reads tests, builds executive reporting, mentors your analyst and sits in product leadership meetings.
- Tracking plan and event taxonomy owned, not outsourced
- Test calendar with a decision log for every test
- Monthly readout you can forward to the board
-
Offer B
ML Decision Model
from $30,000 fixed
6–10 weeks. $30,000–50,000 fixed, or base plus a share of measured lift.
One decision, one model: pricing, segmentation, churn, demand forecasting or delivery and ops economics. Scoped around a dollar outcome and a break-even.
- Data assessment, model, validation and deployment plan
- Plain-English readout the sponsor has already seen before the meeting
- Lookback after launch to check the model held in the field
-
Offer C
AI Automation Build
from $12,000 fixed
4–8 weeks. $12,000–25,000 fixed, 50% deposit. Included in the retainer if you have one.
One automated analytics pipeline: ingest, classify or model, report, alert. Typically text classified into a taxonomy and a dashboard, or a recurring report nobody should be building by hand.
- Monitoring on every run and handoff documents your team can use
- Not included: chatbots, custom LLM apps, automation of work that is not data or reporting
Pricing within a range depends on the number of product teams, data sources, whether a warehouse exists and whether there is an analyst to mentor. Net-15 terms on every invoice.
Need something smaller?
Fixed-price services for businesses of any size, from $500 a month.
A starter dashboard, a monthly scorecard, a sales snapshot, an ads check, a tracking audit, one A/B test read properly, a segmentation or a forecast. Fixed scope, fixed fee, and each one says plainly what is not included. Nothing on the list? Tell us the decision and get a quote within two business days.
Four outcomes from past work
Prior employment, anonymized. Every number was reported to the executive who asked for it. All case studies
About
US-based, remote.
Liftpoint Analytics is run by a product analytics lead who works with companies that have the data but not the senior person to turn it into decisions. Five-plus years in analytics and data science; the last two-plus running a product analytics team for the Chief Product Officer, then the Chief Marketing & Analytics Officer, of a publicly traded two-sided marketplace for high-value physical goods. Before that, analytics at a large transportation infrastructure operator, where ML forecasting improved forecast accuracy about 10%.
How the work runs. Every engagement names the decision it supports, the owner and the date it is needed; analysis that does not change a decision is cut. The tracking plan, event taxonomy and QA come before dashboards, and dashboards come before tests. Every test gets a hypothesis, a powered design, guardrail metrics and a decision log. Every readout is one page a CPO can forward.
I will tell you what the data cannot answer before I tell you what it can.
Retainers are capped at three clients; project work is open to businesses of any size.
- Certifications
- AWS Certified Machine Learning – Specialty
- Tools
- SQL, Python (pandas, scikit-learn, XGBoost, PyTorch), BigQuery, Snowflake, dbt, Tableau, Power BI, Fullstory, GA4, Pendo, Statsig
- Capacity
- Retainers are capped at three clients; project work is open to businesses of any size.
- Code
- Public ML demo on GitHub
Book a free 30-minute Decision Audit
Ten questions, three observations, live. You leave with a recommendation: the Diagnostic, nothing yet, or not a fit. No slides.
Prefer to pick a slot now?
Book directly on the calendar. The form above is optional if you do.
A decision, not a dashboard.
Two days a week, three-month minimum. You get a tracking plan, a test calendar and a monthly readout you can forward to the board.