Results

Every survey has a Results page that combines the responses from every channel the survey has run on.

​

The tabs

The page has four tabs: Summary, Questions, Audience and, when the survey has variants, Variants.

​

Summary

• The numbers at the top: Responses with the daily trend and the count of the last 7 days (7 calendar days in your time zone, today included, whatever the date range), Completion rate (the average share of questions each respondent answers), Median completion time, Peak day, Collecting since and Last response.
• AI reading: a reading of the survey’s responses for the selected period, with What changed and What to do. Each line carries the number behind it. The reading is about the people who answered, never about the survey itself.
• Improve this survey: a separate list for the survey itself, such as questions respondents skip, answers that fit no theme and weeks when fewer responses arrive. It appears only when there is something to fix.
• Signals: what the AI read in every response in the date range, such as sentiment, urgency, customer risk and type. Selecting a number opens those responses in the inbox.
• Themes: the survey’s open-text answers grouped by your organization’s themes, each with its size and how it moved. A theme removed in Settings → Analysis keeps the answers filed under it and is marked (removed).
• Key questions: the questions that say the most, with a link to all of them.

​

Questions

Every question has a card with views to switch between. Most questions have Chart, Over time and Table. Open text has Answers, Themes and Over time. File uploads, screenshots and signatures have Gallery, List and Over time. A date question has Chart and Table. The card remembers the selected view. The chart depends on the question type:

• NPS: the score, then promoters, passives and detractors on one line, for all responses and for each device.
• Ratings: the average with the stars, each rating as a bar, and the share of satisfied ratings (the CSAT), such as “80% rated 4 or higher”.
• Customer effort: the average, and the answers split from hard to easy.
• Sliders and numbers: the median and the range the middle half of the answers falls in, over the whole scale.
• Emoji: the faces in order from unhappy to happy, each with its share.
• Choices: a donut of the answers. When respondents could select more than one answer, the shares can add up to more than 100%, so the card shows bars instead. The 12 most picked options are listed first, and and N more lists the rest. What people typed into Other follows, the newest first, and Show every typed answer lists all of it.
• Yes or no, consent and calls to action: the share that said yes, accepted or clicked.
• Rankings, matrices and points allocation: the average place, a grid of answers per row and the average points per option.
• Dates: a calendar of the dates people selected.
• Open text: every answer in the filtered results, each with how it reads (negative, neutral or positive) and its theme, with a search and the Theme, Sentiment and Attachments filters above the list. The search and the filters look through all the answers, not only the page on screen. An answer written beside a screenshot or a file is listed too, with what was attached to it. Themes lists the question’s eight biggest themes, and and N more lists the rest.
• Contact details and uploads: every contact or file in the filtered results, with a search; uploads also show as a gallery that loads more files as you scroll.

When the survey has answers to questions it no longer has, such as a deleted question or one removed with its variant, Removed questions follows the cards. Usertick keeps those answers and lists each removed question with how many people answered it and its 10 most common values as they were stored, because the wording and options were removed with the question; and N more lists the rest. The page’s filters apply to them. Over time follows the date range: By day for up to a month, By week for up to six months and By month beyond that. In the rare period that holds more than 300 different answers to one question, the chart counts the 300 most common and says so under it. When a date range is selected, each card also compares with the period of the same length right before it: the headline number shows its change, and tables add a vs previous column. All time has no earlier period, so it shows no comparison, and neither does a range that your plan’s history covers only in part. Selecting an answer in a chart or a row in a table filters the whole page to the people who gave it. Tables show ten rows at a time, with pages below, and a longer page scrolls inside the card. A choice question’s charts and tables show its 12 most selected options and say how many more there are.

​

Audience

Who answered shows countries, devices, languages and the pages people answered on, each with its share of the responses. Answers given on the survey’s link page count as one Usertick link page row. When the survey has an NPS, rating or effort question, every group also shows its score next to the overall score, and Scored by selects the question that scores the groups. A dot marks a group that scores clearly below the overall score once at least 20 of its people have answered, and the line at the top names the lowest groups. Selecting a group filters the page to it.

​

Variants

When the survey has variants, this tab compares each variant with the selected group (Everyone else by default):

• A sentence at the top states what the comparison found, based on the numbers.
• One card per group shows who it is shown to, its main score and how far that is from the compared group, its result, its responses and its median time.
• Every question has its own line, with one column per group and the filters Differ, Collecting, No clear difference and Can’t compare. Opening a line shows the full comparison.

A difference counts as clear at 95% confidence once both groups have at least 30 answers. Until then, a group reads Collecting with how many of the 30 answers it has. Better and Worse are used only where the question has a better end (NPS and ratings); elsewhere a clear difference reads Answers differ.

​

About the AI reading

Usertick writes the reading in the background, on every plan, including Free, and it doesn’t count against any AI allowance. The period picker offers Daily, Weekly, Monthly and Yearly. Each reading is written when its period closes, from that period’s responses, and compared with the period before. Periods follow the UTC calendar, and weeks run from Monday to Sunday. Yearly covers the year so far, from January 1 through the last month that has ended, so Usertick writes it again at the end of every month and compares it with the same months of the year before. A period needs at least 30 responses for a reading. A period in which nothing stood out says so, and so does a period whose reading couldn’t be written. The reading about the people who answered and the list about the survey are kept apart on purpose. A line such as “this question was answered by 4% of people” is precise and easy to act on, and in a shared list it would push aside findings about the product, such as “people still export everything to a spreadsheet”. The reading at the top is therefore only about the respondents, and anything about the questionnaire goes under Improve this survey. The reading uses what each response was read as (see How responses are read): themes that grew or that people sound far more negative about, urgent responses taking a larger share, people ready to leave, or a kind of feedback such as bug reports growing. Each of these is compared as a share of the responses read, so a busier month is not mistaken for a worse one.

​

Filter the results

Everything on the page follows the filters. Filters holds the Date range and narrows the page by Country, Device, Language, Page, Ran out of time, Speeders, Time taken, Variant and the answer to any question, and every card and chart updates together, including the themes and the counts on each question. Last 7 days, Last 30 days and Last 90 days count calendar days in your time zone, today included. Ran out of time is offered when the survey has a time limit. Country, Language and Page list the values the survey received in the date range, up to the 250 most common, with a search box; the answer filter of a choice question lists every option the survey has. When the page is filtered to one country, the themes are that country’s themes, counted over that country’s answers. The AI reading and Signals don’t follow these filters. The reading covers every response in its period, and Signals covers every response in the date range, so both show Not filtered when another filter is on. Exports follow the same filters, so a download matches the page.

​

Save a view

Views, at the top of the page, saves the date range and the filters under a name. A view never holds a variant, and it applies to whichever survey is open. Three built-in views, Last 7 days, Mobile and Speeders, are available to everyone and can’t be removed. Views are shared with the whole organization and work as they do in the inbox.

​

Keep the numbers current

Beside Views, a line shows when the numbers were last updated, and its menu offers Refresh now. On the Pro plan and above, the numbers update as new responses arrive and the line reads Live. Turn off live updates, in the same menu, keeps the numbers as they are until the page is refreshed.

​

Related pages

• Read and act on individual responses in the Inbox.
• Combine several surveys and custom charts on a Dashboard.
• See how the scores are calculated in Metrics.