Skip to main content
Narrow the research GetWhys draws from — by company size, industry, seniority, function, region, or recency. Filters control which interviews GetWhys can see when it answers you. Set them once and every question you ask after that draws only from the slice you defined. Use them when a general answer isn’t precise enough: you’re selling into enterprise but the answer reads mid-market, you need the last six months rather than the last two years, or you want the practitioner’s view instead of the executive’s.

​
Why filters matter

GetWhys answers from a corpus of interviews spanning industries, company sizes, and roles. Relevance ranking does most of the work — ask about enterprise IT leaders and you’ll mostly get enterprise IT leaders. But “mostly” isn’t always enough. When you’re building a segment-specific campaign, a competitive teardown for one vertical, or a persona refresh, you want a guarantee rather than a tendency. Filters are that guarantee. There’s a second reason, and it’s the one that changes how people prompt. Typing “focus on enterprise buyers” into your question is a request — the model interprets it and does its best. Setting the enterprise filter is an instruction — GetWhys can only retrieve from interviews that match. Both are useful. When precision matters, use the filter.

​
How to apply a filter

Cmun6bc8h000a0agmet1zgqzx (1)
  1. In any chat, select “Filters” on the bottom-left of the message input.
  2. Expand any category from the menu on the right-hand side, and select the values you want.
  3. Ask your question.
Your filters stay active for the rest of that conversation until you change them.
Filters are not retroactive.A filter applies only to questions you ask after setting it. An answer already on screen was drawn from the unfiltered corpus, and adding a filter won’t change it.If you want the filtered version of a question you already asked, re-ask it. Copying and pasting the same prompt is fine — that’s the fastest way to compare a broad answer against a narrow one.

​
What you can filter on

Company size

Employee count bands. Useful for separating enterprise from mid-market and SMB buying behavior.

Industry

Restrict to the verticals you sell into. Buying criteria and vocabulary shift a lot between them.

Seniority

Practitioner, manager, director, VP, C-level. The executive view and the practitioner view are often genuinely different answers.

Function

The department the interviewee works in. More reliable than job title, because titles vary wildly between companies.

Perspective

The seat the interviewee was in — decision maker, influencer, user, implementer, tool owner. Use it when the question is about a stage of the journey rather than a segment.

Region

Geography of the interviewee’s company. Worth setting when regulation or local competitors shape the answer.

Date range

How recent the interview is. Use this when you’re working on a fast-moving category and older interviews would muddy the picture.
You can also choose which source types GetWhys draws from — the GetWhys research corpus, your Organization Data, or both. Unchecking GetWhys data gives you an answer grounded only in your own uploaded calls and research. Unchecking Organization Data gives you the outside-in view alone.

​
Start broad, then narrow

Ask your question with no filters first. Read the answer and the citations. Then decide what to constrain.
This ordering matters more than it sounds. Two reasons: You’ll find useful things you would have filtered out. An interview with someone who doesn’t look relevant on the surface — different industry, adjacent role — often contains the sharpest quote in the answer. Filtering first means never seeing it. Stacked filters shrink the pool fast. Each filter is a constraint on the same corpus. Enterprise and financial services and C-level and last six months and EMEA may leave very few interviews, and a thin answer built on three interviews is worse than a strong answer built on forty. GetWhys will tell you when the data is thin, but it’s easier to widen from a good answer than to diagnose a bad one. The habit worth building: broad question → check citations → add the single filter that matters most → compare.

​
Confirming a filter took effect

The sources at the bottom of a chat should display the metadata — job title, industry, company size, interview date — you selected. If you see something that doesn’t match your filter, tell us. Contact [email protected] with a link to the conversation.

​
Filters vs. asking in the prompt

Both work, and they’re good at different things. Use them together. Filter to the segment, then use the prompt for the nuance inside it.
Some things you might want to filter on aren’t filterable — company founding date, tenure in role, whether someone is a former employee of a specific vendor. Describe those in the prompt instead.If GetWhys doesn’t have the interviews to answer, it will say so rather than guess, and you can file a Research Request with those criteria in the screener.

​
When a filtered answer comes back thin

That’s a signal, not a failure. It usually means one of three things:
  1. Your filters are too narrow. Remove the least important one and re-ask.
  2. The corpus is thin on that specific slice. File a Research Request — describe the audience precisely, including the same criteria you were filtering on. New interviews are delivered on your AnswerSLA and become searchable from every future chat.
  3. The recency window is too tight. If you filtered to the last six months and the category moved recently, there may not be enough recent coverage yet. A Research Request is the fix here too.

​
For PMM and marketing leaders

Filters are what make a research answer defensible in a room. “Buyers care about X” invites debate. “Across 40 interviews with VP-and-above IT leaders at 5,000+ employee financial services companies in the last six months, here’s what they said” ends it — and the citations are right there if anyone wants to check. They’re also how one corpus serves a whole team without everyone getting the same generic answer. Your enterprise PMM and your mid-market PMM can ask an identical question and get two genuinely different, equally well-grounded answers.