Roles we recruit
Seven practices, 27 functions. Start at the practice when you cannot yet name the title; go straight to the function when you can.
Browse all 229 roles Brief us on a seatFlat 15% of first-year base · shortlist in 10 working days
A data scientist builds the models that predict: lead and account scoring, churn and expansion propensity, pricing and forecasting. Notebooks are easy to show, so we screen for models that reached production, a decision one changed, and how they explain a result to someone who does not read code.
A data scientist’s value shows when a model changes a decision, not when a notebook looks good. Each stage below has a signal that predicts it, and we test for that signal before you meet anyone.
The ramp, and the test behind each milestone
Weeks 1 to 3
The warehouse, the CRM and the history behind each field, plus conversations with sales, marketing and customer success about the decisions they make today. Agrees one question worth a model, and the baseline it has to beat.
On track looks like one written target, such as which accounts convert or churn, with the current method as the baseline.
How we test for it before you meet them
A scoping exercise in our screen: a described business problem and a data dictionary. We score the question they choose, what they would measure and what they would ask us for.
Weeks 4 to 6
Clean features, a simple model first, and an honest comparison with how the team decides today, tested on data the model has not seen.
On track looks like a first model that beats the baseline on held-out data, with the gain explained plainly.
How we test for it before you meet them
We ask for a model they built and how they tested it: the split, the baseline and what they found when it failed. A former manager confirms it.
Weeks 7 to 10
The score or forecast delivered where people work, in the CRM, the planning sheet or a dashboard, with the logic written down so a sales or marketing lead can trust it.
On track looks like a model output that a team uses in its weekly work, with an owner for each decision it touches.
How we test for it before you meet them
We ask which of their models is still used today and by whom. Models that lived only in notebooks are where we dig.
Weeks 11 to 13
Monitoring for drift and data breaks, a check on whether decisions improved, and a short note on what the model cannot do.
On track looks like a monthly check on the model's accuracy and use, and one decision measurably better.
How we test for it before you meet them
A reference call with a business owner about a model they shipped: whether it changed a decision and how they handled it when it was wrong.
After day 90
Strong data scientists start saying no to questions a spreadsheet can answer, and yes to the next one that justifies a model, with the data work it needs.
On track looks like a short roadmap of models ranked by value and data readiness that leadership agrees.
How we test for it before you meet them
We ask which question they would model next in your business and why. A specific answer tied to revenue or cost predicts the scientist who picks well.
Five responsibilities. If the seat is mainly dashboards and reporting, it is an analyst; if it is mainly pipelines and data quality, it is a data engineer.
Lead and account scoring, churn and expansion propensity, with a baseline they have to beat.
Revenue, pipeline and demand forecasts, and price testing, beyond what a spreadsheet can handle.
Setting up tests and holdouts so the business knows whether a change worked.
Delivering outputs in the CRM, planning tools or dashboards where decisions are made.
Plain statements of what a model says, how sure it is and where it should not be trusted.
Three neighboring seats absorb work people assume belongs here. Drawing the line wrong is how a data science hire ends up building dashboards or fixing pipelines.
That seat ownsCampaign and funnel reporting, attribution views and the weekly numbers the team runs on.
Hire it instead whenthe need is clear reporting and answers from existing data, not a predictive model.
That seat ownsThe pipelines, warehouse models and data quality that analysts and scientists depend on.
Hire it instead whenthe data is late, duplicated or untrusted, so no model could be trusted either.
That seat ownsDashboards and self-serve reporting for leaders across the business.
Hire it instead whenleaders cannot find their own numbers and the questions are descriptive.
The title stays the same; the work does not. Pick the one that matches your data and your decisions.
Builds scoring, propensity and forecasting models for sales, marketing and customer success.
Designs tests and pricing analysis, and tells the business what changed and why.
The first data hire in a company, doing analysis, simple models and some data engineering.
The questions we ask in our own screen. Use them in yours; the warning signs are as useful as the answers.
Every candidate we put forward is scored on these five, and the scores travel with the shortlist, gaps included. Four out of five with a named weakness beats a clean sweep you cannot verify.
| What we score | Weak evidence | Strong evidence |
|---|---|---|
| Models in use | Notebooks and prototypes that never reached a workflow. | Models that a team still uses, with the decision they changed. |
| Choosing the question | Starts from technique, not from the decision. | Picks a question tied to revenue or cost and names the baseline. |
| Testing against a baseline | Reports accuracy on training data. | Held-out tests against how the team decides today, in business terms. |
| Explaining results | Explains the method instead of the meaning. | States what it predicts, how sure it is and what to do, in plain language. |
| Fit with your data | Built for clean research data, with no experience of messy CRM history. | Has worked with data like yours: your CRM, your volume, your gaps. |
Public US benchmarks, read October 2026. Salary.com’s band and Built In’s self-reported median of $120,000 agree closely for a mid-level data scientist. Seniority is the large step: Built In’s senior data scientist median base is $140,000, and the BLS band is wide because one occupation covers every industry and level.
Each bar runs from the 25th to the 75th percentile, the middle half of reported pay.
Base salaryWages including commission (BLS)Median base salaryAverage total pay, base plus variable
Data Scientist
Salary.com base, Built In total
$108,704 to $128,376 base$145,852 average total
Senior Data Scientist
Built In, self-reported, the next seat
$140,000 median base$176,357 average total
Business Intelligence Analyst
Salary.com base benchmark, neighboring seat
$102,090 to $123,502 base
Data scientists
BLS, all industries
$85,660 to $158,880
Data Scientist and Business Intelligence Analyst bars: Salary.com base salary, 25th to 75th percentile, United States, as of October 1, 2026. Rings and the Senior Data Scientist dot: Built In median base salary and average total compensation (base plus additional cash, self-reported), reviewed October 13 and 14, 2026. Data scientists: U.S. Bureau of Labor Statistics, Occupational Employment and Wage Statistics, May 2025, annual wage at the 25th and 75th percentiles (SOC 15-2051), all industries. Public aggregates, not ZaphyrRecruit placement data.
A data scientist uses statistics and code to answer questions that predict or explain. In a go-to-market team that means models for lead and account scoring, churn and expansion risk, pricing and forecasting, plus the tests that show whether a change worked. The job also includes getting results into the tools where people decide, and explaining them plainly.
Salary.com puts a data scientist at $108,704 to $128,376 base, with a median of $118,416 (October 2026). Built In’s self-reported figures show a $120,000 median base and $145,852 average total pay, rising to a $140,000 median base for senior data scientists. The BLS band for all industries runs from $85,660 to $158,880. Seniority and domain move the number most.
A data analyst describes what happened: reporting, dashboards and answers from existing data. A data scientist predicts what will happen or tests why: models, forecasts and experiments, with more statistics and code. Many teams need an analyst first. If leaders cannot find their own numbers, hire the analyst; if the numbers are clear and decisions need prediction, hire the scientist.
When you have years of reasonably clean data, a decision that a model would make better, and someone who will act on the result. Without those, the hire spends the first year on data cleanup. If the data is messy or scattered, a data engineer or analyst usually pays off sooner. We will say so on the first call.
Statistics, Python or R, SQL, and judgment about which question deserves a model. For go-to-market work, add experience with CRM and marketing data, experiment design and clear communication. Tool lists matter less than evidence of a model that reached production and a decision it changed. Ask to see one in use today.
Give them a described business problem and a data dictionary, and ask what question they would model first, how they would test it and what they would need from you. An hour is enough. It shows judgment and baseline thinking, which polished notebooks hide. Avoid long unpaid projects: strong candidates decline them.
Ten working days from a signed brief to a shortlist of three to five calibrated data scientists, each with our scoping exercise scored and a shipped model checked with a former manager. Interview speed decides the rest: good data scientists rarely run one process at a time, so a decision within two weeks of the first interview keeps them.
Tell us the decision you want to improve, the data you have and who will use the result. We come back within one business day with how we would run the search and what the seat pays.
We build shortlists from people doing this work now, often before a seat is advertised. Set up a job alert for data roles, or send your CV and we will come to you when a seat opens.
The data scientist is one of seven seats in this function. If the gap sits elsewhere, start there.