StatQuestions FAQ

Question 1

Our feedback lives in a system you have probably never seen. Can you still bring it in?

Accepted Answer

Yes. Setting up your ingestion pathway with you is part of how we work, not an add-on. On the 30-minute data-mapping call we look at where your feedback lives (a case system, a phone system, shared inboxes, exports) and agree the way in. We have connected to systems we had not seen before; the fit is assessed on the call, not from a list on a web page.

Question 2

What do you need from us to get started?

Accepted Answer

One person who knows where the feedback lives and can grant access or send an export, and about 30 minutes on a call. For a manual export you can be live the same day. For a connected system it usually takes a few days, most of which is us, not you. Card numbers and personal details are masked before anything is saved.

Question 3

Is there a cap on how many emails, calls or tickets you will read?

Accepted Answer

No. We do not cap interactions and we do not charge per response. Rules handle the volume first, AI handles the residual, and your team sets the rules on what is left, which is what keeps large volumes affordable. Voice and email are scored on the same scale so one root cause is counted once across both.

Question 4

We have strict data governance rules. Can we use StatQuestions?

Accepted Answer

Most teams remove or anonymize PII before uploading, customer name and email address add no analytical value. The text body, date, and metadata fields are what the analysis runs on. If your legal team requires a data processing agreement, we provide one. Customer text is processed and not retained for model training.

Question 5

Our customer text is locked in a vendor platform with no export. Can you still help?

Accepted Answer

Usually, yes. If the platform has an API or a scheduled export we can build the pathway with you. If it allows neither, the first step is to fix that contractual limitation, and we will tell you that plainly on the demo call rather than sell you a workaround.

Question 6

Can we use StatQuestions without sending data to a third-party cloud?

Accepted Answer

Private cloud and on-premise options are available for enterprise accounts with strict data residency requirements. Contact us and we'll discuss the right deployment model. For most regulated industries (healthcare, financial services), our standard deployment with a signed DPA is sufficient.

Question 7

We already have a survey and experience management platform. Why do we need StatQuestions?

Accepted Answer

Survey and experience management platforms are measurement tools. They tell you scores went up or down. StatQuestions tells you why, with statistical significance, and what it costs. If your current platform can hand you a prioritized fix list with dollar impact attached to each root cause, you probably don't need us. If it gives you dashboards and you're still debating what to fix, that's the gap we close.

Question 8

We have an internal data science team. Can't they do this?

Accepted Answer

Possibly, but at what cost and speed? A data science team building a custom ticket classification and waste quantification system typically takes 3–6 months, costs $200K+ in team time, and produces something they have to maintain forever. StatQuestions does this in days and doesn't require their involvement. The better question: what else could your data science team be doing if they didn't have to build this?

Question 9

We already have a BI dashboard with ticket volume trends. Is that not enough?

Accepted Answer

Volume trends tell you that escalations are up 18%. They don't tell you which failure pattern is causing 34% of those escalations, what that pattern costs in rework, or what to fix to reduce it. StatQuestions picks up where BI leaves off, it reads the text, not just the counts.

Question 10

We use a large language model API internally to summarize tickets. Why do we need StatQuestions?

Accepted Answer

Summarization is a different task than statistical pattern detection. A summary of one ticket is useful for an agent. A statistically validated pattern across 40,000 tickets, with chi-square proof, business unit attribution, and cost quantification, is what justifies a capital investment. StatQuestions isn't a summarization layer; it's an analysis system.

Question 11

We're below 5,000 contacts per month. What should we do?

Accepted Answer

Wait, or use StatQuestions for a focused single-question analysis on a specific issue type or time period. Below 5,000 contacts, patterns aren't statistically separable from noise at the company level. Above 5,000, they are. If you're growing toward that threshold, book a demo now and we'll tell you exactly when you'll be ready.

Question 12

We're a pure transactional B2C company with no retention model. Are we a fit?

Accepted Answer

Probably not for the core product. The causal analysis tier requires an outcome you care about, churn, repeat contacts, refunds, escalation rate. If customers transact once and never return, there's no retention signal to attach root cause to. You might still get value from Tier 1 issue classification, but the ROI story is weaker.

Question 13

CX at our company reports to Marketing and doesn't have operational authority. Does that matter?

Accepted Answer

It matters a lot. StatQuestions produces root cause findings that require operational decisions, process changes, policy edits, system fixes, headcount reallocation. If the person using StatQuestions doesn't have a seat at the table where those decisions are made, the findings will sit in a slide deck. The right owner is VP CX, COO, VP Quality, or Head of Operations, someone whose job depends on fixing the problem.

Question 14

We're mid-procurement on a large platform deal that includes text analytics. Should we pause?

Accepted Answer

Ask the platform vendor what their root cause and waste quantification output looks like specifically, not the feature demo, but an actual prioritized fix list with dollar impact per issue type. If they can show you that clearly, it may be redundant. If they show you a dashboard or a sentiment score, there's still a gap. We're happy to sit on a call with your evaluation team.

Question 15

I'm not a statistician. Will I be able to use this?

Accepted Answer

Yes. StatQuestions runs the tests for you and writes the result in plain language, "this moved, and it is real" or "this moved, and it could easily be chance." You don't prepare data by hand, write a query, or choose a test. And if someone on your team does want to check the working, the method, the sample size and the confidence sit on the same screen as the answer, so the finding is never a black box.

Question 16

How quickly will we see results?

Accepted Answer

Tier 1 (issue classification and pattern dashboard) is ready within 24–48 hours of a clean file upload. Tier 2 (waste quantification) adds 1–2 hours of configuration for handle time and labor rate inputs. Tier 3 (causal correlation to outcomes) requires joining an outcome file, typically a half-day of data prep, and the analysis runs in minutes once uploaded.

Question 17

How do we know the root cause analysis is accurate and not just an AI guess?

Accepted Answer

Every root cause finding is grounded-ds-control in the most statistically frequent patterns, not a random sample and not a language model's inference about what "probably" caused something. The 5 Whys runs against text evidence that actually appeared at volume. Chi-square tests validate that patterns are statistically distinguishable from noise (p < 0.05 threshold). You see the supporting verbatims at every step. If a pattern has 12 cases, we flag it as inconclusive rather than reporting it as significant.

Question 18

What if our data quality is poor, lots of short or vague ticket descriptions?

Accepted Answer

Lower text quality reduces classification precision. If 40%+ of your tickets are one-sentence descriptions like "billing issue" or "can't log in," the taxonomy will be less granular. We can still find patterns in those cases, they just map to broader issue types rather than specific root causes. We'll do a data quality assessment on the demo call before you commit.

The questions worth asking before you buy.

Getting your data in

Data & Security

We already have…

Fit & Volume

Results & Proof

The three qualifying questions

Something missing?