NextFluent

Questions

What research teams ask before they bring a file.

Short answers. Where a longer one exists, the link takes you there. Anything missing, write to hello@nextfluent.com.

Files

Which files can I add?

An SPSS .sav file, which carries the question wording, the answer labels and any user-missing codes, so nothing has to be typed in. Or a CSV with one row per person and one column per question. A CSV with no labels of its own can be joined by a two-column labels file, column name and question wording, and by a pairs file naming the look-alike questions.

What about a weight?

A weight column in the file is applied and named on the card. The weighted figure leads and the unweighted one stands beside it. Significance tests on weighted figures run on the effective base. How weights are counted.

How do I add a second wave?

On the Waves tab, name the wave and drop the next file. Each wave stays its own file. A question with no column on a wave has no figure on that wave, and the Waves tab lists which questions were asked on both, dropped, or added. The Debrief tab then says what moved since the wave before, tested and with receipts, and what the file can count about why.

Do I upload by email?

No. You upload in the desk, over HTTPS, to the EU. We do not accept files by email, and we ask you not to attach one when you write to us.

Asking

How do I phrase a question?

The way your client says it on the phone. “How many switched supplier last season?” binds to the question about switching supplier, in the past tense. Add “by region” to cut it. Add “among those who changed supplier” to narrow it. Ask for “the mean” or “the average” on a scale and you get the mean. When a study is new on the desk, the Ask screen offers a few questions worked out from its own codebook, so you can see what it reads before you type.

Can I ask in Dutch or French?

When the file is worded in Dutch or French, yes. The desk reads the question wording and answer labels as they are, and the Ask screen says so: “Ask this study in English or Dutch”. An English question binds to “Hoe tevreden bent u” with the same receipt as a Dutch one. On a file worded in English, ask in English.

An English question finds a Dutch or French question through a built-in word list, the same kind of fixed file as the English dictionary. It knows fewer words than a person does, so on a Dutch or French study the desk asks you to choose more often at first, and each choice is kept. Words that are particular to your study, a product or a place, are best taught on the Questions tab.

Why did my question come back with no figure?

The card says why. The question names something the study never asked about. No field in the study is a clear match. Two fields match too closely, or a word in your question is not settled, in which case you are asked to choose. The question asks about a group, a time or a place the survey did not, or asks two things at once. Or the question is in the study but was not asked on that wave, in which case the card names the look-alike that was asked and the percentage it would have given. The refusal rules.

What does “You choose” mean?

Your question came nearest to one survey question, but a word in it is not that question’s own. It may belong to a different question, or to nothing in the study. The desk names the word, shows the questions it could mean, and counts nothing until you pick. You pick once: the choice is kept for that wording, and for a cut of it. Every word settled.

Do I have to use the survey’s exact words?

No. A built-in dictionary settles a word that means the same as one of the question’s own: “closest” for “nearest”, “hold” for “have”. It is a fixed list of English words that ships with Fieldlock, not a language model, and nothing you type leaves the desk to look a word up. A word the dictionary does not tie to the question waits for your choice, once. You can also teach a question your client’s words on the Questions tab.

Can I make a study less strict?

Yes, study by study, on its Pins tab, and separately for your desk and for client pages. “Every word settled” is where every study starts. “Other questions’ words settled” asks you only when a word belongs to a different question. “Rules and dictionary only” asks you nothing, and is for the desk only. A figure counted under anything but the strictest setting says so on its receipt.

What is a pin?

Your decision, remembered. When two questions score too close together, or a word in your question is not settled, the desk shows the candidates and asks you to pick. Pin one and every later ask with that wording goes to it. The card says it was pinned, and pins are listed on their own tab.

Can I ask across several studies at once?

Yes, from the studies page. Each study answers on its own tile, with its own waves and its own receipt. Two studies are never blended into one percentage.

The numbers

What does the hash prove?

That the figure on the card came from that study, that wave, that question, that field and those rows, counted without a language model. Run the same question on the same file later and the hash matches. Change a row and it does not. Copy the receipt from the card and anyone can re-hash it on the desk, without a login. What is in a receipt.

Where do the column letters come from?

Every column in a cut gets a letter. A cell lists the letters of the columns it is higher than at the 95% level, tested with a two-proportion z for shares and a Welch z for means, on effective bases when the figures are weighted. The footnote under the table says which bases were used.

How are scales and “don’t know” handled?

A 1–5 or 0–10 question is read as a scale. The card gives the favourable box, top 2 on five points and top 3 on 0–10, and the mean, both on the people who answered on the scale. A “don’t know” sits outside the scale and is left out. The reading band then tells you what the figure would be with the don’t-knows in the base, and whether that changes the picture.

What does “Asked only of the 484 who …” mean?

The desk read the routing back from the file. When a question was put only to people who gave a particular answer earlier, the pattern of blanks shows it, and the card says who was asked, in a sentence. How routing is read.

What can I ask for beyond a share?

A named box: “top 3 box on satisfaction”, “bottom box”. A net across the items of a grid: “aware of any of these brands”, “heard of none of them”, or two answers added up, “Flanders or Brussels”. A band on a number: “spend over €50 a week”, “between 40 and 80”, “aged 50 or more”. A ranking question: “ranked first”, “in the top 2”, “mean rank”, with every item in a table under the figure. Each of these takes “by region” and “among those who”, and each writes what it did into the receipt, so the hash covers the box, the net, the band or the view as well as the field. With two or more waves the card adds a table of the same figure on every wave, with the change wave on wave and whether it is real at 95%. Any card comes out as a CSV with its hash in the last row.

Can it tell me what drives satisfaction?

Yes, and it is the one card that is not a plain count. Ask “what drives overall satisfaction” and the desk takes the rated items in the file, correlates each with the outcome on the people who answered both, and shows the share of the explained variation each item carries, with its correlation, how many pairs it rests on, and whether it clears 95%. It is Pearson on the codes, nothing more elaborate, and the card says so. The receipt carries the specification: the outcome field, every item with its pairs and its correlation, the method and the floor on pairs. Run it again on the same file and the hash matches. A client page does not run it; the desk does.

Can it run a segmentation or a conjoint?

No. Fieldlock counts. A segmentation, a conjoint, a regression or a factor solution is a model, and a model needs choices the desk should not make for you. Run it in the package you already have, write the result back into the file as a column, a segment per respondent or a utility band, and bring that file. The desk reads the column like any other question. “Switched supplier by segment” then cuts by your segmentation, with letters between the segments and a receipt over the rows, and your client can ask it on their page.

Does a language model write any of this?

No. Every figure is arithmetic on the rows of your file, and every sentence on a card is built from the file’s own wording. Every receipt records model=none. Your response file never goes to ChatGPT or any other model provider. The one place a model may read: if the owner of your desk switches the reader on, you can give a study the questionnaire it was fielded with, and a model reads the wording off it so the card carries the question as the respondent saw it. It sees the questionnaire and the codebook’s labels, never a row, and you accept what it proposes.

Has anyone checked that a chat really gives the 25%?

Yes, on 9 November 2026, on the public Harborline sample. Handed the Next season file and asked what percentage delisted last season because of the hygiene score, the largest open model we ran gave 25% in four runs of four, in prose noting that the question was not in the file and answering anyway. The smaller models gave 61.5%, 50%, 30%, 12% and 0%. Every answer, the script and the two files are on the Compare page.

Your client

What does my client see?

One page with your team’s name on it. On it, the cards you chose to send, each with its figure, wording, base and receipt. If you open a study to them, a box where they can type their own question and get an answer under the same rules. Never the file, never the codebook. See a client page.

What if my client words it differently from the survey?

A word that means the same as the question’s own is settled by the built-in dictionary and they get the figure. A word nobody has settled gets no figure: their page says the wording is not settled yet and that their research team has it. The question appears at the top of your Ask screen with a Settle button. You choose once, and it answers on their page from then on. A client is never shown a figure counted on a guess.

Do I see what they asked?

Every question your client types is listed on your client page desk, with what came back and when. If one came back too close to call, you can pin it from there.

How does my client get in?

By the link you send them, the way you already talk to them. A new page opens only for the people you name: you list their work emails, and each signs in with a link sent to their own inbox, so a forwarded link opens nothing. You can untick that on a page, and then anyone who has the link can open it. A listed address also finds every page it has been given on the brand side, from every team that uses Fieldlock. A new link closes the old one, and you can close the page at any time.

I have several clients. Does each get their own page?

Yes. Each client page has its own link, its own studies, its own listed addresses and its own cards. Nothing from one page shows on another. You make a page from the Client pages screen, put the studies on it that this client may see, and send the link. Removing a page takes its link, its cards and its addresses with it; the studies stay on your desk. Each client after the first is €1,200 a year, agreed in writing before the page opens.

The free sample

What does the homepage sample do?

It runs the real engine on Harborline, a made-up food packing study with two seasons. Your question goes to our server in Frankfurt and the card comes back; nothing else about you is sent with it. Ask it anything. You will see a counted card, a refused card and a receipt, exactly as they appear on the desk.

Can I try my own file?

On the desk, not on the website. Book a 30-minute call and we open a desk with your file on it, behind your sign-in, in Frankfurt. Do not email the file.

What does the free version not do?

It does not keep your studies between visits, it does not remember pins, and it has no client page. Those live on the desk.

The desk and pricing

What does it cost?

The desk is €2,400 a year for one seat with every study you run. With a page for one client it is €3,600 a year. Each further client is €1,200 a year. A colleague on the desk is €600 a year, agreed in writing before they sign in. Pricing.

Can a colleague work on the desk?

Yes. From the Account page, the owner of the desk types a colleague’s work email and the desk sends them a link. They choose a password, set up an authenticator app, and from then on they sign in on their own. They see the same studies, client pages and log, and every card they send or question they ask is written in the log in their name. The owner can take them off the desk again from the same page; their login stays, with nothing to open. A colleague is €600 a year, agreed in writing before they sign in.

How do I start?

Write to us from your work email. The desk is invite only, so we open a login for you and set up your first study with you on a call, in an afternoon. Do not send the file by email; you upload it in the desk.

Where is my data, and who processes it?

Database, sign-in and file storage are on Supabase in Frankfurt. The application is hosted on Vercel. Nobody else processes your response files. If the owner of your desk switches the questionnaire reader on, the text of a questionnaire you give a study goes to Mistral AI in the EU for that reading, and nothing else does. Security and data.

Still a question?

Write to us. We answer from Antwerp, from a person who built the desk.