Sharing & launching

Study statuses

Transitions (on the Launch step): Draft → Use your audience → Live; Live → Pause → Paused; Paused → Go live again; any state → Close study; Closed → Reopen as draft (back to an editable draft).

The study funnel. Every study moves through the same steps, left to right: Build (title, cards, categories) → Audience (who takes part) → Screener (qualifying questions) → Pre-launch (review reach, price and readiness, nothing is spent) → Launch (go live) → Results. When you're buying responses, the ordering and payment happen on these same steps, not on a separate screen.

Readiness checks before going live

To launch you need:

• A title.
• At least 2 cards.
• For closed/hybrid sorts, at least 1 category.

The participant link

On the Launch step you'll find the participant link in the form …/s/XXXXXXXXXX (a unique 10-character code). Click Copy and share it however you recruit: email, Slack, your own panel, a QR code. (When you buy responses from Sorted Research you don't need this link, we handle recruiting for you; see below.)

• The link only works while the study is Live. In draft, paused, or closed states participants see a "not available" or "closed" message.
• Custom/branded slugs aren't available yet; the code is auto-generated.

Recruiting through a panel (User Interviews, etc.)

Panels and recruitment tools track who completed your study by passing a unique code through the link. SortedResearch supports this out of the box.

1. In your recruitment tool, use your normal participant link (…/s/XXXXXXXXXX) as the task URL. The tool tacks a tracking code onto it, e.g. …/s/XXXXXXXXXX?iid=pp_123456.
2. Copy the redirect (completion) link the tool gives you, for User Interviews it looks like https://www.userinterviews.com/study-task-events/custom_link_with_redirect/complete?iid=TRACKINGID.
3. Paste that link, exactly as given, into Redirect URL on completion in your study's Settings. Leave the word TRACKINGID in place.

When a participant finishes, SortedResearch reads the tracking code from the link they arrived on and swaps it into the redirect in place of TRACKINGID, so the tool credits the right person automatically. Nothing else to configure per study.

Buying responses from Sorted Research

The easiest way to get participants is to buy responses from Sorted Research: tell us who you want to hear from and how many responses you need, and we find and recruit a matched audience for you. This is the default way to recruit on every plan, you don't have to recruit anyone yourself, paste anything anywhere, or manage a panel. It works the same for a card sort, a pairwise study, or a Best–Worst study. You pay per order, before it goes live.

Choose where participants come from, the Audience step

You set this on the Audience step of your study:

• Buy responses from Sorted Research (the default). We find and recruit the right respondents for you. You choose who they are and how many you want, and we handle the recruiting end to end.
• Bring your own respondents, free. Publishing a study costs nothing: choose "Use my own participants" under the audience controls, publish on the Launch step, and share the link with people you recruit yourself (a panel, your own list, social, a QR code). You only ever pay when you place a panel order; running a study on your own link is free on every plan.

A study uses one or the other, never both. If it mixed them, people arriving through your own link would use up the responses you paid for, and your order could never fill. Once a study has collected a response the setting locks, because that response came from one source. Set up a new study if you need the other one. Opening your own link to check it doesn't lock anything, only a completed response does.

Who can take part, audience targeting

On the Audience step you describe who you want to hear from, for example age, gender, income, region, or whether they have children. We match your study to real people who fit. As you set this up you'll see two live numbers:

• How many people match your audience (your reach).
• The price per response for that audience.

The narrower or rarer the audience, the fewer people match and the higher the price per response, a very general audience is cheap and plentiful, a very specific one costs more. How long the study takes also affects the price: longer studies cost a little more per response (you set the expected length on the Audience step). You never enter any technical rates; you just pick who you want and we work out the price.

What it costs, a clean price per response

Sorted Research shows you a single, clean price per response (for example "$10.00 per response"), chosen automatically from your audience. Your total is simply that price × the number of responses you want. There are no hidden add-ons and nothing else to configure.

For a very large or very specialised audience there may be no instant price, instead you'll see "request a quote." Click it and we'll get you a custom price by email. Everything else works the same once the price is agreed.

The price can move as you change your audience or your screener: a tighter audience or extra screening questions mean fewer people qualify, which can raise the price per response. Whatever you finally see is exactly what you pay.

Pay before it goes live

You pay for your sample before the study starts recruiting. The flow on the Launch step is:

1. Place the order. This sets the study up but doesn't spend anything or recruit anyone yet.
2. Pay. You're taken to a secure Stripe checkout for price per response × number of responses (e.g. "Sorted Research, 200 responses"). Nothing recruits until this is done.
3. Go live. Once payment clears, the Go live button unlocks and recruiting begins.

If you change the number of responses or the audience after paying, we re-check the price and ask you to pay any difference before it can go live, so you're never launched on a stale price. A study waiting on a custom quote is handled by us and can't be self-paid until the quote is agreed.

If your order doesn't completely fill, automatic refund

Sometimes an order delivers fewer responses than you paid for (for example a niche audience runs dry). When you close the order, we compare what was actually delivered against what you paid for and automatically refund the difference, you only ever pay for the responses you actually got. There's nothing to claim; the refund goes back to your card. So if you close an order early, you're refunded for every response that wasn't delivered.

Quality control on bought responses

Sorted Research checks every bought response for quality the moment it finishes, and acts on it for you. There is nothing to approve and no queue to work through.

Clearly bad responses are rejected automatically. Someone answering faster than the questions can be read, clicking the same side of the screen almost every time, contradicting themselves on a repeated question, or dumping the whole task into one pile, is rejected the moment they finish. A rejected response never counts toward your target, never affects your scores, and is not paid for. Our recruiting partner is told immediately, which frees the slot so they send a replacement. You do nothing.

Borderline responses are flagged for you, not hidden from you. Some responses are not bad enough to throw away but are worth your eye: more contradictions than most people in your study, noticeably faster than everyone else, a rating scale that does not make sense. Those are not rejected, because rejecting an honest-but-noisy respondent loses you real data. Instead your participant list is ranked, with the ones worth a look at the top, each showing in plain English why it is there, so you can check the claim rather than take it on trust.

You decide what to do about those. If you read one and want it gone, remove it: that rejects it with our recruiting partner, refunds it and frees the slot for a replacement, exactly as an automatic rejection does. If you are happy with it, do nothing, it already counts.

You get the quality control of a managed panel without having to babysit it, and you keep the final say on the judgement calls.

How your target is counted

Your target responses counts the responses actually delivered to you. A rejected or removed response frees its slot again, so a replacement respondent can take it and you still end up with the full number you ordered. The same applies to any demographic quotas you set in the screener, so a replacement isn't turned away by a slot a rejected response was still holding.

We never over-deliver, and you're never billed past your target.

This section applies to studies where you buy responses from Sorted Research. A study you recruit yourself counts every completed response, exactly as before.

Your order pauses itself when it's full

When your order reaches the number of responses you paid for, it pauses itself and stops recruiting. You don't have to watch it or stop it by hand, and you're never billed past your target.

A paused order isn't finished, it's just resting. Your study stays open and your results stay exactly as they are. On the Launch step the button changes to Resume recruiting, and the order panel says the order is paused.

Deleting responses usually gets you replacements

If you review what arrived and remove some responses, the slots they were using open up again, up to the number you originally paid for.

Removing a bought response tells our recruiting partner not to count or pay that person, which is what frees the slot on their side so they can send a replacement. That happens in the background, usually overnight.

Recruiting restarts by itself about an hour after your last removal. You do not have to press anything. The results screen tells you the time it will happen, so it is never a surprise.

The hour is deliberate, not a delay for its own sake. Our recruiting partner only accepts a rejection while your order is paused, so if recruiting restarted the instant you removed your first response, none of your remaining removals could be refunded. Waiting until you have stopped means every removal is refunded and replaced. Each further removal restarts the hour.

So remove everything you want gone in one sitting. Once recruiting is running again you have to pause the order before you can remove anything else, which the Launch step lets you do at any time.

If you would rather not wait, press Resume recruiting on the Launch step yourself and it starts immediately. Doing that, or pausing or closing the order by hand, also cancels the automatic restart, because your own choice always wins.

When a removal can't be refunded

There are two cases where removing a response does not get your money back and does not get you a replacement. We tell you before you confirm, either way, and the response's data is removed from your results regardless, so your findings are unaffected.

The response was already settled. Our recruiting partner settles each response once and only once. Almost always that happens when you remove it, which is what refunds it. But if the respondent's browser failed to report back to the partner when they finished, we settle it for them straight away so they are paid for work they genuinely did, and the partner will not settle it a second time.

This is rare, and we keep it rare on purpose: we only settle a response early when it is clean on every quality check. Anyone carrying a quality signal, which is exactly who you are likely to remove, is deliberately left unsettled so that removing them still refunds and replaces them as normal.

The order is closed. Closing an order is final, and our partner only accepts a rejection while an order is paused. So do your reviewing and removing before you close, and while the order is paused. A full order pauses itself, which gives you that window without having to arrange it.

Pausing and resuming an order yourself

You can stop and restart recruiting whenever you like, from the Launch step of your study:

• Pause recruiting stops new people being sent to your study. Anyone already part-way through finishes normally and still counts.
• Resume recruiting starts it again, and only ever fields up to the number you already paid for. Resuming never costs more than your original order.
• Close study is final. Closing settles any shortfall refund, and a closed order can't be reopened.

Always pause and resume from inside SortedResearch. It keeps your study and your order in step with each other.

If buying is temporarily paused

Occasionally we may pause buying across the whole platform (for example if there's an issue on our recruiting side). While it's paused, no new study can start buying, and any order already in field stops letting new people in, nobody is left stranded mid-study.

Your data and your money are safe either way. Anyone already part-way through finishes as normal, responses already collected stay exactly as they are, and your responses to review panel keeps working so you can still approve or reject held responses. Your study settings stay editable. When buying is switched back on, your order carries on filling.

Who's active right now

The dashboard header shows a live "active now" count: how many participants are mid-study at this moment across your studies, refreshed about every 15 seconds (hover for the per-study breakdown). A participant counts as active while they're on the screener or sort and has been seen in the last ~90 seconds. It's mainly a heads-up before making changes, so you don't disrupt anyone in the middle of sorting.

Editing a live study

While a study is live you can freely edit: participant instructions, the thank-you and study-full messages, redirect URL, consent text, privacy URL, target responses, and the screener. You can also add or change cards and categories, but be aware that changing them mid-study can make responses harder to compare.

Two things lock once a study has responses. The sort type (open/closed/hybrid), because changing it would change what the sorts you have already collected mean; and where participants come from (buying responses from Sorted Research versus your own participant link, covered above), because those responses came from one source. Both can still be changed on a live study that has no responses yet, and opening your own link to check it doesn't count, only a completed response locks them.