Support Center

It scores a real sales or customer call against the skills that make calls succeed. You import a call transcript; the app works out the call type, then grades the rep on eleven selling skills — opening, discovery, problem development, objection handling, next steps, and more — against a standard you (the manager) set. You get one clear score with a written explanation and a list of next steps, instead of a gut-feel opinion.

It doesn't record calls or listen in live. You give it a transcript that already exists, and it gives you a read on that one call. See Getting Started for the full walkthrough.

Project Moneyball is the company behind STUDIO, founded by Avner Baruch. Its mission is to help revenue teams “Scale Faster, Better & More Effectively” — turning real calls into objective, coachable signal instead of opinion. STUDIO is its call-scoring product.

Learn more at projectmoneyball.com.

Sales and customer-success leaders and managers who coach reps, enablement teams measuring whether training moved the needle, and reps who want objective feedback on real calls — not opinions.

The Home page asks you to pick a path, and you can switch anytime from the left menu:

• Skill-Based Diagnostic — score one call. This is where most people spend almost all their time.
• Bespoke Playbook — compare two cohorts of calls (your top performers against your core reps) and turn the gaps into a coaching playbook.

If it's your first time, start with a single call on the Skill-Based Diagnostic.

The menu collapses to an icon strip using the chevron at its top.

Four steps on the Skill-Based Diagnostic page:

• Import a transcript — click Import transcript, or drop the file onto the Message Center.
• Check the setup — Call Details fill in automatically; review the expected minimum and weight for each skill.
• Run — a quick Sanity Check runs first, then each skill lights up as it's scored.
• Read the result — the score, the analysis, and the recommendations. Download PDF gives you a clean branded report to share.

The full walkthrough is in Getting Started.

A plain-text transcript: .txt, .vtt, .srt, .md, .log, or .csv, up to about 5 MB. Export the transcript from your call recorder or notetaker and import the file here.

Word documents and PDFs won't work, and renaming one to .txt won't help — the app checks the contents, not the extension, and will tell you it doesn't look like a plain-text transcript. Export as text instead.

200,000 characters — comfortably more than a long call. If a transcript runs past that, it isn't rejected: the call is scored on the first 200,000 characters and the rest is ignored. For a very long recording (an all-day workshop, several calls stitched into one file), split it and score the parts separately so nothing meaningful is cut off.

The first run on a new transcript takes a couple of minutes. After that, re-running the same transcript is quick and free — the saved analysis loads instantly — until you change a setting, which scores it again.

A brief pre-flight pass that confirms the transcript actually looks like a real call before scoring begins — so you don't spend a run on a meeting agenda, a support chat log, or a file that came out garbled.

If it passes, scoring continues as normal. If it can't be scored you'll see “Transcript not analyzable” with the reason, and the run stops. If the transcript is usable but thin, you'll get a warning and scoring continues.

We only show the rep's role when the call gives enough evidence to determine it reliably. If it can't be determined, we leave it out rather than guess — and a note appears in the Message Center explaining why.

On calls with more than one rep on your side, the name and role shown are the call owner — the rep who actually drove the conversation and was scored, not whoever opened the call.

Not yet — it's on our roadmap. You'll see an Integration page in the left menu marked Roadmap; that's where it will live. Today you export the transcript from your notetaker and import the file.

STUDIO is designed to sit on top of the recorder you already use, so there's nothing to rip out. If you tell us which tool you're on (info@projectmoneyball.com), we'll factor it into what we build first.

The score is built around 100: 100 means the call met exactly the standard you set. Above means it beat that standard, below means it fell short. The scale runs to 120.

More detail in Getting Started → Reading the score.

Under the score you'll see something like “72% — Medium Confidence”. Coverage is how much of the scorecard the call actually gave evidence to score — not every skill comes up on every call. Confidence follows from coverage: High at 85%+, Medium from 60–84%, Low below 60%.

Low coverage doesn't mean a bad call. It means there was less to go on, so read the score with that in mind.

That's the safety gate doing its job — it stops a good average from papering over a real problem. A score is held back when:

• A critical skill lands far below its bar — capped at 74.
• Two or more critical skills are below target — capped at 79.
• Coverage is under 60% — capped at 79, with confidence set to Low.

A capped score shows in red as a flag that the call needs a closer look. It's a signal, not a glitch.

A separate page (in the left menu) that shows how the eleven skills connect — not just how each one scored. After you run a call it tells the connected-skills story: why each skill matters, how one feeds the next, and, when a call falls short, the single root cause to fix first so the whole chain improves with it — instead of a flat list of low scores.

It reads your most recent run, so there's nothing to configure — score a call and the page fills in. More in Getting Started → The Cross-Skill Diagnostic.

Alongside the score, STUDIO runs a separate MEDDPICC qualification scan and reports, for each of the eight elements, whether the call showed strong, weak, or no evidence. Buyer-stated evidence always counts for more than anything the rep merely asserts.

It never factors into the final score — by design. Qualification isn't a one-time event and can't be judged from a single call: a discovery call legitimately won't have a paper process yet, and a renewal call won't re-establish pain from scratch. Scoring one call against the full checklist would punish it for things it was never meant to cover.

The right way to measure qualification is to collect evidence across every touch in the lifecycle and keep a score that updates as the relationship progresses. That's something Project Moneyball can help you build — get in touch.

Download PDF gives you a clean, one-page branded report you can save or send. The Copy button on the Score Analysis and Recommendations boxes copies that box as plain text, for pasting into your notes, a one-to-one agenda, or an email.

A scorecard is your definition of good for a given call type — the expected minimum and weight for each of the eleven skills. Open the Scorecards page from the left menu, pick a call type, set your preferences, and save.

From then on it applies automatically whenever a call of that type is scored — there's nothing to load each time. Any call type you leave as Default uses STUDIO's system settings, and Revert to default undoes your changes. You can also save the current setup straight from the Diagnostic page.

Expected minimum is the bar, from 1 to 5. The skill's score is measured against it: meet it and the skill is on target, fall below and it counts against the call. Set the bar to 0 to make a skill optional — doing it adds a small bonus, skipping it costs nothing.

Critical does one extra thing: it acts as a safety gate. If a critical skill badly misses its bar, the whole call's score is capped — so a strong call can't paper over a critical failure.

Yes. Your scorecards are private to your account and aren't visible to anyone else.

The app identifies the call type from the transcript and loads the matching scorecard. Fifteen are built in, spanning the whole lifecycle:

• Cold Call / Prospecting Outreach, Inbound First Call, Intro Call (first meeting / qualification)
• Discovery, Product Demo, Business Case / Value Review, Technical Validation (POC / POV / Security)
• Follow-up (mid-cycle), Commercial / Negotiation
• Customer Onboarding, Customer Adoption / Health Check, Business Review / QBR / Renewal, Expansion / Upsell, Escalation (support / business)
• Default / Fallback — used when the call type can't be identified confidently

The Lifecycle Stage row highlights where the identified call sits, from Prospecting through to Expansion.

It answers the question every leader has: what do my best reps actually do differently? You give it two cohorts of calls of the same call type — your top performers and your core (or struggling) reps. STUDIO scores every call, averages each skill per cohort, and shows the two side by side, sorted by the biggest gap.

Then it turns those gaps into a coaching playbook: the handful of skills that separate the groups, what the top performers do on each, the single thing most likely missing, and concrete actions to bridge it. You can download the whole thing as a PDF.

Four steps on the Bespoke Playbook page:

• Choose the call type and the scorecard to judge against.
• Upload a .zip of transcripts for Top performers.
• Upload a second .zip for Core / struggling reps.
• Click Generate playbook.

Each cohort takes 2 to 10 transcripts — aim for 5 to 10 a side for a result worth acting on. Keep the cohorts clean: mixing weaker calls into the top-performer zip is the fastest way to a playbook that says nothing.

One .zip per cohort, containing up to 10 plain-text transcripts (.txt, .vtt, .srt, .md) — all of the same call type. The upload box takes a zip only, not loose files. Anything beyond the 10-transcript limit is left out, and you'll see a note saying so.

Roughly 1.5 minutes per file — so a full set of 20 can take up to about 30 minutes. Keep the tab open while it runs: closing or reloading it stops the run.

And yes — every transcript in the two zips is a scored call, so a 20-file playbook uses 20 calls. Calls you've already scored are cached, so re-running them is fast and free.

Go to the app and enter the email address you were invited with. You'll get a one-time PIN by email — paste it in and you're through. There's no password to remember or lose.

Sign-in is handled by Cloudflare Access, and access is currently invite-only: named users, invited by us. If your PIN doesn't arrive, check your spam folder first, then email us.

Not self-serve yet. During early access we invite users by email address, one account per person — send us the addresses at info@projectmoneyball.com and we'll set them up. Shared team accounts and seat management are on the roadmap.

Sign out sits at the bottom of the left menu and ends your session. Next time you'll sign in again with a fresh one-time PIN.

Two different things:

• Paused — an administrative hold. Nothing is deleted: your scorecards, credits, and results are all intact, and resuming puts everything back exactly as it was.
• Blocked — this happens when a file you imported contained content our input guard flags as malicious (for example, instructions hidden inside the transcript designed to hijack the scoring engine). Imports stop at that point.

Either way, the way back is the same: email info@projectmoneyball.com and we'll look into it with you. If it was a false positive on a legitimate transcript, we want to know — that's useful signal for us.

Every account starts with 5 free diagnosed calls. It's a count, not a clock — there's no trial timer running down, so you can take your time and score them when you have calls worth scoring.

No per-seat fees and no subscription — you pay for the calls you choose to score:

Prepaid credits don't expire. Full comparison on the pricing section of our site.

We'd rather tell you that plainly than send you to a checkout button that doesn't exist. In the meantime, if you run out mid-evaluation, just ask — we're not in the business of stopping an interested team from finishing their look.

One transcript scored, once. Re-running the same transcript without changing anything is served from the saved analysis — fast, and it doesn't cost another call. Change a setting and it scores again, which does.

The Bespoke Playbook scores every transcript in both zips, so a 20-file playbook uses 20 calls.

We built the product around data minimization. We do not store your full transcripts — each transcript is processed in memory for a single scoring run and is never saved at rest or written to logs. We keep only the derived analysis (scores, confidence, short evidence excerpts, summaries, recommendations) so your results can be re-displayed without reprocessing the call.

Your transcript is processed in memory by our Cloudflare Worker and sent to Anthropic (Claude) to score the call; the results are assembled, validated by an output gate, and returned to you. What we keep lives on Cloudflare:

• Derived analysis in Cloudflare KV, keyed to a one-way SHA-256 hash of the transcript, kept ≤ 30 days then auto-deleted.
• Your scorecards in Cloudflare D1, under your account.
• Usage counters against a hashed email.

Data is encrypted in transit (TLS) and at rest. Access is identity-gated via Cloudflare Access (Zero Trust) — every request carries a signed identity token we verify before any processing. Hosting is on Cloudflare's US / global edge. See the Security Brief.

We use a small set of sub-processors to run the service, each under a data-protection agreement. We don't sell your data and don't use it for advertising.

• Cloudflare — hosting, database, key-value storage, and identity / access.
• Anthropic (Claude) — AI scoring of the transcript. Transcript data sent via the API is not used to train their models.
• A transactional email provider — only if/when you use the email-a-report feature.

We give notice before adding or replacing a sub-processor. Details in the Data Handling Overview.

Yes — though you don't have to, since we never store your transcript in the first place. If your team is under strict privacy or compliance rules, the Anonymize page (left menu) gives you a self-serve prompt: paste it into any AI assistant — including a local/offline model for maximum privacy — add your raw transcript, and it returns a synthetic copy with real names, quoted pains, and figures swapped for consistent stand-ins, while the call's structure and every behavioral gap are preserved, so it scores the same as the original. Import that copy instead.

It's an optional extra layer for the extra-cautious, not a fix for a gap. More in Getting Started → Anonymize your transcript.

Yes. Because full transcripts aren't stored and the derived-analysis cache auto-expires within 30 days, most call data is short-lived by design. Scorecards and account data are deleted on request or account closure, and you can request a fresh run that ignores and overwrites the cache. Email info@projectmoneyball.com.

Yes. A formal DPA is available for execution before any paid use, and we're happy to complete security questionnaires and share our sub-processors' compliance reports (Cloudflare, Anthropic — both SOC 2 audited) on request. Email info@projectmoneyball.com.

The authoritative detail lives in our Trust Center: the Privacy Notice, Security Brief, and Data Handling Overview.

A red box in the Runtime Monitor means that one skill couldn't be scored on this run — usually a brief, temporary hiccup with the scoring engine, not a problem with your call.

The app handles this for you. It retries automatically a couple of times — you'll see a “retrying automatically” note and the box reset and re-run. Most hiccups clear by themselves.

If a skill is still red after those retries, click Run once more (only the red skills re-run; the rest load instantly from the saved analysis). If it keeps coming back red, it's genuine: that skill simply didn't come up in the call, or the transcript is too short or unclear for it. It's left out rather than guessed — and if several drop out, coverage falls and you'll see a Low-confidence note, so try a fuller transcript.

The scoring engine is briefly unreachable, or you've hit a throttle. Nothing is broken on your side — wait a little and Run again.

There are two technical limits, well above normal use: 20 requests a minute and 100 calls a day per account. They're there to keep the service healthy for everyone, and they're separate from your credits. If you're legitimately bumping into them, tell us — that's a conversation worth having.

Run only lights up once a transcript is loaded. If it's greyed out, import a transcript first. If a run seems stuck, click Stop and re-import.

The Message Center at the top of the app explains what's happening at each step — read the latest message there first; it usually says exactly what's missing.

The Sanity Check decided the file isn't a call it can score, and it tells you why. The usual causes are a file that isn't really a call transcript (an agenda, a chat log, notes), a transcript too short or garbled to read, or an export where speaker turns were lost.

Re-export from your notetaker as plain text with speaker labels intact, and try again. If it's genuinely a real call and still gets rejected, send us a note — we'd like to see that one.

They're visible in the left menu but not active yet — open either and you'll see a short description and nothing else. That's deliberate: we'd rather show you where we're going than hide it.

• Buyer-Side Diagnostic — the other half of the call: buyer engagement, sentiment shifts, buying vs. risk signals, and decision-readiness.
• The CRO Lens — team-level rollups: performance trends, coaching priorities, and deal-risk across your reps.

• Integration — connecting your notetaker (Gong, Fathom, Fireflies…) so calls score automatically, instead of exporting and importing.
• Self-serve billing — buying credits without emailing us.
• Team accounts — inviting your own users and managing seats yourself.

Priority is shaped by what people actually ask for, so if one of these is blocking you, say so — it counts.