How to Track Customer Satisfaction Over Time
min read - Updated on Nov 15, 2026You can track customer satisfaction over time in Kimola by keeping the feedback of one channel in a Space and adding a new file on a regular rhythm, such as every week or every month. Kimola classifies every file with the same custom models, so a theme means the same thing in the first file and in the latest one. After each upload it recalculates the Space's NPS from what customers write, runs statistically tested Signals that show which topics rose or fell, and writes what the latest file changed. When your team acts on a finding, Kimola reads the same feedback again weeks later and tells you whether customers noticed. This guide follows the path from the first Space to the first measured Task.
Tracking satisfaction in a Space is available in the Enterprise package. It is also available in Kimola On-Premises.
The Goal
Satisfaction moves slowly and from many directions. A new release, a price change, a slow month for deliveries or a fix in the help center can each change how customers write, and a team that reads one export at a time sees a series of snapshots, each with its own themes and its own numbers. The question this guide answers is the one a head of customer experience, a product leader or a quality team asks again every month: are customers more or less satisfied than before, which topics explain the change, and did what we changed make a difference?
Kimola measures satisfaction from the text of the feedback itself, in three ways that answer three different questions. The Space's NPS says in one number, from −100 to 100, whether its feedback leans positive or negative. The Satisfaction Score and the Dissatisfaction Score rank the themes customers are most and least satisfied with, adjusted for the size of each theme. The Dimensions of your Funnel, such as Pain Points, say what lies behind a movement. All three are calculated from the text of reviews, tickets, conversations and open-ended answers, so they work in channels where no customer gave a score. The NPS you know from surveys is the score of a 0–10 question and is a different measure; see How NPS Is Calculated in Spaces.
A channel that keeps producing feedback belongs in a Space, because every new file is read against the records before it, and a Report analyzes one dataset once. If you have a single batch to study, such as one survey wave, a Report is the better fit. See Reports vs. Spaces.
For a food delivery app, a Space of app store reviews might show an NPS of 14 after its first file and 6 three months later. The Trends list shows late deliveries rising, and the team that runs dispatch gets a Task. Weeks after the Task is done, Kimola tells whether customers mention late deliveries less often.
What You Need
You need the Enterprise package or Kimola On-Premises, and a few things ready before the first step. Each is easier to get right at the start than to change later.
A channel to follow. Choose one channel of feedback that you can export as a file again and again, such as app store reviews, support tickets, survey answers, call transcripts or social media posts. Kimola reads .xls, .xlsx, .csv and .tsv files, one file per upload, up to 30 MB each, and each row becomes one record. The feedback of one channel goes into one Space.
A date for every row. The first file needs one column with the date of each piece of feedback. Every file of the Space is placed on the same timeline by this column, which is what makes the trends possible.
A Funnel with a sentiment classifier and a content classifier. The Space's NPS is calculated from the sentiment classifier of its Funnel, and a Space whose Funnel has none has no NPS. The content classifier gives the themes the Space follows, and the Satisfaction and Dissatisfaction Scores need both classifiers. Dimensions are optional and answer why. Step 1 explains how to choose them.
Descriptive columns (optional). Columns such as the product, the region, the plan or the survey score are kept with every record. Kimola doesn't classify them, but you can filter and compare by them, and Signals can compare the groups they form.
Teammates. A Task can only be assigned to a team member, so the people who will own the findings need to be members of your team.
Step 1: Set Up a Space
A Space rests on three decisions: the Funnel that classifies its feedback, the name of the channel it follows, and the schema its first file sets. Settle them before the first upload. Every later file is read the way the first one was set up, and satisfaction can only be compared over time when every file is read the same way.
Start with the Funnel. A Funnel is the set of custom models every Space of your team classifies feedback with, and it holds at most one model of each kind. For this guide, add a sentiment classifier, because the Space's NPS is calculated from it, and a content classifier, because its labels become the themes Kimola ranks and follows. Build the content classifier around the topics customers raise in this channel, in the words your teams already use. Add Dimensions when you want to know why satisfaction moves: Pain Points asks what goes wrong or creates difficulty, Unmet Needs asks what is missing or desired, and Underlying Causes asks why it occurs. Leave Enable Multi-label Classification on when one piece of feedback often covers several topics. A Space keeps the Funnel it was created with, and editing a Funnel later applies only to the files uploaded after the edit, so settle the labels of the content classifier before the first file goes in. If your team already has a Funnel with a sentiment classifier and a content classifier that fit the channel, use it. See Create a Funnel, Create a Custom Model with a Prompt and Multi-Label vs. Single-Label Classification.
For a software company, the content classifier of a support-ticket Funnel might have labels such as Onboarding, Billing, Integrations and Reliability, with a Pain Points model next to it. Every ticket is then read through the same themes, so a rise in Billing next quarter is a rise in the same topic.
Editing a Funnel later applies only to the files uploaded after the edit, so settle the labels of the content classifier first.
Then create the Space. Open SPACES in the main menu and click New space, choose the Funnel, and name the Space after its channel, such as NPS Survey or Social Media. A channel name stays right for as long as the Space lives, while a name such as "Q3 export" goes out of date with the next file. The optional description holds up to 250 characters and is the place to say where the feedback comes from and how often it is uploaded. If you follow more than one channel, give their Spaces the same Funnel, so that they share their themes and sentiment, their NPS can be read side by side and Home shows all of them together. Click Create Space. See Create a Space.
[Screenshot: The Details step of the new Space wizard, with a channel name and a description filled in]
The new Space is empty until its first file. Open it and click Upload the first file. This file sets the Space's schema, the structure every later file follows. In the Schema step, choose the Date Column, the column with the date of each piece of feedback. Choose the date the feedback was written, because the Space's trends follow this column and the date of an export would put every record of a file on the same day. Then name the Text Fields, the columns Kimola classifies, for example Review or Comment. When one text is spread over several columns, click Join another column to classify them as one text. Every other column of the file is kept with each record and is not classified, and these are the columns you filter and compare by later. A column that is not in the first file's schema is not kept from later files, so include every column you may want to use, such as the product, the region or the plan. If your files identify the customer behind each row, you can also choose the Customer Column; it is optional and can be added later. See Add Files to a Space for the whole wizard.
[Screenshot: The Schema step with the Date Column chosen and a Text Field named Review]
A column that is not in the first file's schema is not kept from later files.
Step 2: Upload Files on a Schedule
Satisfaction can only be followed while feedback keeps arriving, so the second step is a routine: export the new feedback, upload it and let Kimola classify it. Choose a rhythm that fits the channel, such as one export every Monday for a busy channel or one at the end of every month for a slower one, and keep to it. Each upload is read against the records before it, and a steady rhythm keeps the trends easy to read. A rhythm that gives every upload enough records to compare serves you better than a frequent one: a channel with a few dozen records a month is better followed with a monthly file than with a weekly one.
Export only the feedback that is new since your last upload. Each row of a file becomes one record, so a row that appears in two files is counted twice and inflates the Space's volume and trends. Keep the column names of the export the same from file to file.
Open the Space, go to Files and click Add data, then choose the file in the Dataset step. Every file after the first uses the Columns step in place of the Schema step. Kimola lists the Space's columns and matches them to the columns of your file by name, and it remembers the matches you choose, so a regular export that keeps its column names is matched without asking again. Check the matches and click Continue. The SUMMARY shows the file's rows and the queries it will use, counted as rows × models × Dimensions, so a file with many rows or a Funnel with several Dimensions uses more queries. Click Upload. If the Space already has a file with the same name, Kimola warns you before it counts the same export twice. Tick It is a different file, upload it anyway only when it is a new export that happens to have the same name, as a weekly export often does.
[Screenshot: The Columns step of a later file, with the schema's columns under SPACE and the file's columns under FILE]
Queries are counted as rows × models × Dimensions. See How to Calculate Queries, and kimola.com/pricing for the queries in each package.
Kimola then classifies every text field of every row with every model of the Funnel. On the Files page the file shows Waiting to be classified, then Classifying with its progress, then Classified, and you can leave the page in the meantime. A file counts in the Space only once it is classified. Every record keeps the date from its own Date Column, so feedback from last month that you upload late is read as last month's feedback. If a wrong file went in, take it back from the Files page with Take this file back: its records leave the Space, and its NPS and Signals are recalculated.
You don't start anything else. When a file is classified, Kimola recalculates the Space's NPS, runs its Signals again over all the records and writes its Insights again where the evidence moved. The first Signals and the first Insights appear after the first file.
A wrong file can be taken back with its records. See Take Back a File from a Space.
Step 3: Follow NPS and Trends
After every upload Kimola brings the Space up to date, and the results are spread over a few pages of the Space, each answering one question. Read them in this order after each file: how customers feel overall and in each theme, how this file compares with the one before it, what the latest files changed, and which findings passed a statistical test.
Read the NPS and the Themes Behind It
Start on Overview, the first page of the Space. The NPS card shows how customers feel as a ring on a scale from −100 to 100. The ring fills green clockwise when the NPS is positive and red counterclockwise when it is negative, so you can read the direction before the number. Hover over it to see how the number is made, for example "NPS -8: 7% Positive − 15% Negative". Kimola recalculates it every time a file is classified, from all the records of the Space. Every file adds to the records before it, so the NPS of a large Space moves slowly. Feedback that is neither positive nor negative, such as a plain question, counts in the total but in neither share, so what matters most is the direction the NPS takes from one upload to the next.
The themes behind the number are on the same page. In ALL THEMES, every category and label shows its NPS with a two-sided bar, and a theme's NPS reads how customers speak about that theme, so one record can be positive about one theme and negative about another. BEST PERFORMING THEMES and WORST PERFORMING THEMES list the three themes customers are most and least satisfied with. They are ranked by the Satisfaction Score and the Dissatisfaction Score, which are adjusted for the size of a theme, so a theme that hundreds of customers complain about ranks above a handful of angry records. The Performance page shows the scores of every theme, and the dates next to its Filters card score the themes of one period, such as the last month, so you can compare it with the period before. NPS is a difference between two shares, so it swings widely when a theme has few records, and the number of records is shown next to every theme. See Overview in Spaces and Performance in Spaces.
[Screenshot: The Overview of a Space with the NPS card, BEST PERFORMING THEMES and WORST PERFORMING THEMES]
The NPS of a large Space moves slowly. To see what one upload changed, compare files.
Check how many records a theme's NPS reads before you act on it.
Compare One Upload with the Next
The Space's NPS reads all its records, so it describes the Space as a whole. To compare periods directly, open Compare in the menu of the Space. Click Files in the Compare card, choose 2 to 5 files, for example this month's upload and last month's, and click Compare. The option is offered once the Space has at least 2 classified files. The page shows every file as a column: its volume and its trend, its sentiment as the number and share of Positive, Negative and None records, its themes, its Dimension answers and its most frequent words. Reading across the columns, you see whether negative feedback grew from one file to the next and which themes carry the change.
The labels of an opened classifier follow the first column. A theme that a later file has and the first one doesn't is listed under Mismatches at the end of that file's column, and it is often the most telling result of the comparison. Pick the earlier file first, so that the themes that appear only in the latest file are listed there. To compare segments instead of files, choose Column values and pick values of a column such as the product or the region. See Compare in Spaces.
For an online course platform, comparing the files of March and April might show the Negative share of the sentiment rising from 18% to 27%, with video playback listed under Mismatches in the April column because March has no such theme.
[Screenshot: The Compare page with two files compared, showing Volume, Trend and Sentiment side by side]
See What Changed with the Latest Files
Back on Overview, the top of the page tells you in words what changed. What your feedback says now is a short text Kimola writes about the Space after every new file. Next to it, With the latest files lists the themes and Dimension answers that are new, rising or falling in the newest files compared with the records before them, and Fading lately lists what customers mention significantly less often now. Each item carries its trend: New, Back again, Rising, Falling or Fading. Kimola decides with fixed rules what is written and how each theme moved, and an AI model only writes the words, so a text changes only where the feedback changed. These texts read the Space's latest records, at least 1,000 records over at least 30 days, or the whole history of a smaller Space, and a line under each text shows its window. See What is an Insight? and How Insights Are Updated.
[Screenshot: What your feedback says now, with the With the latest files and Fading lately lists]
The Dimensions page shows why. For every Dimension of your Funnel it lists the answers customers give with their share, and an answer whose share changed significantly since the last upload carries ▲ or ▼ with the points it moved, for example "▲ 6 pts". Click an answer to see Over time, its share of the records that answer the question, month by month, together with the themes customers talk about when they give it. Kimola calls a change only when enough records answer on both sides: when too few records of the last upload answer a question, the page says so and shows no ▲ or ▼. A Pain Points answer that rises on this page while the NPS falls on Overview gives you a lead on what moved satisfaction. See Dimensions in Spaces.
[Screenshot: The Answers list of a Dimension with a ▲ badge, and the Over time chart of the selected answer]
Read Signals and Records Over Time
Open Signals to read the findings that passed a statistical test. Trends show how the mentions of a topic changed over recent periods, and the periods come from the Date column, grouped into units such as weeks. Differences show where a topic is mentioned more, or more negatively, in one group than in the others. In a Space every file is a group too, so a topic that stands out in the latest upload becomes a Signal even when its share of the whole history has barely moved. Every card gives its numbers, the group it is compared with, a few quotes and Why this is shown. Give each Signal a verdict, Known, New & important, Unimportant or Artifact, with an optional note, and click Open in Content to read the feedback behind it. A verdict follows the same finding into later runs, so each upload's review starts where the last one ended. A new Space can have no Signals yet: when the tests find no group or period that stands apart with enough evidence, Signals stays out of the Space's menu, and more files, or more descriptive columns in them, give the tests more to work with. See Signals in Spaces and Review Signals.
For a mobile network operator, a Trend might show coverage in rural areas mentioned in a larger share of the feedback month after month for a quarter, with the share of negative mentions rising along with it.
[Screenshot: The Signals page of a Space with the Trends tab open and a Signal card with Why this is shown]
The Content page adds a view of the volume. Records over time draws three lines, All, Positive and Negative, so you can see when the volume rose or fell and whether a rise came with more negative feedback. Drag across the chart to focus on a period. See Content in Spaces.
A new Space can have no Signals yet. More files, or more descriptive columns in them, give the tests more to work with.
Step 4: Act and Measure
A finding that nobody owns doesn't change satisfaction. The last step hands a finding to a teammate and then checks whether it worked, using the same feedback you keep uploading.
Make a Task from a Finding
You make a Task where you found the finding. Click Make a task on a Signal card or on a theme in ALL THEMES on Overview, or filter the Content page to the feedback you mean and click Make a task of the feedback these filters show. A window opens with a draft title and description that Kimola writes from the finding, in the language of the Funnel: why it matters, what to do first and how you will know it is done. Edit the draft, choose the Assignee, an optional Due date and the Priority, and add under Keep informed the teammates who should be told when the Task changes. Then click Create the task. If a teammate has already made an open Task of the same finding, the window shows it first, so you can follow it instead of starting a second one. If Jira is connected for your team, the Task can be sent there as well. The Task keeps the finding's evidence, with its numbers and its NPS, as it was the day you made it. See Create a Task for every place a Task can start and Send a Task to Jira.
[Screenshot: The Make it a task window with a drafted title and description and the Assignee, Due date and Priority fields]
An open Task of the same finding is shown first, so you can follow it instead of starting a second one.
Read Whether Customers Noticed
The clock starts when the Task is marked Done. Two weeks later, and again four weeks later, Kimola reads the same feedback in the files uploaded since and compares it with the weeks before the Task was made. It compares the finding's share of the feedback, which is its records divided by all the feedback of the same weeks, so a busy month with more reviews doesn't make a problem look bigger, and it compares the finding's NPS. The result appears on the Task's page under DID CUSTOMERS NOTICE?: Customers noticed, No change, Got worse or Too little feedback. While an improvement holds, Kimola checks again at eight and twelve weeks. If customers start to talk about the problem as they did before, the Task says "It came back: customers talk about it as they did before the task." and offers Open it again. The results of all the Tasks that were done in a period are together in the Team overview. See How Kimola Measures Task Impact, Manage Tasks and Use the Team Overview.
For a hotel chain's review Space, a Trend shows slow check-in rising, and the front office manager gets a Task. In the four weeks before the Task was made, slow check-in appeared in 9% of the reviews. After the hotels add a self check-in desk and the Task is done, it appears in 5% of the reviews in the next four weeks, while the number of reviews stays about the same. The Task shows Customers noticed.
This is why the uploads of Step 2 matter beyond the trends: the feedback of the weeks after a Task is the feedback its measurement reads. A Task is measured only when each period it compares holds enough records, so a Space that receives one small file a month may need longer before a Task can be measured. Files added more often fill the weeks compared sooner, so Tasks show Too little feedback less often.
More frequent files give fewer Too little feedback results.
What You Get
A satisfaction number you can follow. The Space's NPS, recalculated after every file, and the NPS of every category and theme next to it, so you see the direction of satisfaction and where it moves.
The themes behind it. The best and the worst performing themes by Satisfaction and Dissatisfaction Score, for the whole Space or for the period or segment you filter to.
What changed after every upload. A text about what your feedback says now, the themes and answers that are new, rising, falling or fading, and a side-by-side comparison of any two files.
Why it moved. The answers of your Dimensions with their change since the last upload and their share month by month, so a falling NPS comes with leads on its causes.
Signals. Statistically tested Trends and Differences, each with its numbers and quotes, including the topics that stand out in the latest file.
Tasks with measured impact. A finding becomes a Task with its evidence kept as it was that day, and once the Task is done, Kimola tells whether customers noticed.
Next Steps
Once the Space runs on its own rhythm, Actions in Spaces writes an Action Plan from everything the Space has classified, with up to 12 evidence-checked actions that you can turn into Tasks. If the same customers write in more than one channel, a Customer Column gathers their feedback from every Space on one page. When different teams own different parts of the feedback, the Kimola team can set up Departments with you, each with its own NPS and trend; talk to your Account Manager. To analyze the comments of a survey in the same way, see How to Analyze NPS Survey Comments.