Sales changes: Where is growth coming from?
Compare revenue and order counts by month, region, or product. Separate overall growth from changes in individual segments.
Mockitt AI · Start with a question
Start with Excel or CSV.
Understand what changed—and what to do next.
Give your sales, channel, or operations spreadsheet a clear question to answer. Use the Mockitt AI desktop app to clean up, summarize, and compare local files. Define your metrics and calculations, then organize the results into charts or a report so every conclusion can be traced back to the data.
Mockitt AI desktop app· Windows / macOS · Work with files in the desktop app

An AI data analysis tool helps turn a business question into a concrete analysis task. First, establish what the data can answer. Then decide whether to summarize it, compare groups, or investigate unusual records. The Mockitt AI desktop app starts with file-based tasks, so you can keep asking questions, checking results, and making changes around an Excel file, CSV file, or related materials.
Questions from your everyday work
Compare revenue and order counts by month, region, or product. Separate overall growth from changes in individual segments.
Start by comparing visits and leads, then consider conversions and costs to evaluate each channel. Traffic volume alone does not tell the whole story.
Summarize expenses by department, category, or month to identify where spending changed most.
Organize tasks by owner, status, and time period to see what remains unfinished and how quickly work is moving.

First, get the data on the same terms
Before analyzing anything, ask the app to identify the worksheets, fields, and records it has read, and confirm what each row represents. An order table and a line-item table cannot use the same method to count orders. Likewise, adding monthly totals to daily records will double-count the data.
Start with one specific task
Walk through a channel comparison using six rows of sample data. Switch between the original data, month-over-month results, and findings and actions. You can check every number in the tables.
Read the channel data file I specify. First, list the fields, record count, missing values, and duplicate records. Then summarize visits by month and channel. Do not change the original data. If you cannot read the file or perform a calculation, explain the specific reason before proceeding.
Refine the result Keep the original data and save the summary separately. Do not automatically treat blank values as zero; list the affected records first.
| Month | Channel | Visits |
|---|---|---|
| Last month | Website | 1,240 |
| This month | Website | 1,480 |
| Last month | Paid campaigns | 860 |
| This month | Paid campaigns | 792 |
| Last month | Partners | 310 |
| This month | Partners | 356 |
This month's visits · Matches the original data
Compare this month's and last month's visits by channel. Calculate the percentage change as (this month − last month) ÷ last month, and round it to one decimal place. If last month's value is zero, show N/A and explain why. Include totals for both months. Calculate the overall percentage change from the totals, not by averaging the individual channel percentages.
Refine the result Use the same data to create a channel comparison chart. State the months and metric in the chart title, and keep the summary table so the chart values can be checked.
| Channel | Last month | This month | Change | % change |
|---|---|---|---|---|
| Website | 1,240 | 1,480 | +240 | +19.4% |
| Paid campaigns | 860 | 792 | −68 | −7.9% |
| Partners | 310 | 356 | +46 | +14.8% |
| Total | 2,410 | 2,628 | +218 | +9.0% |
This month's visits · Matches the original data
Write a short analysis based on the verified summary table. Separate facts shown by the data, possible explanations, and information that is still missing. Without campaign cost, conversion, or audience-quality data, do not decide whether a campaign is worth continuing or describe changes in visits as changes in revenue.
Refine the result Add a “What we still need before deciding” section at the end of the report. Distinguish the questions the current data can answer from the data that still needs to be collected.
| Observation | What the data shows | Next step |
|---|---|---|
| Overall growth | Visits increased by 218, or about 9.0% | Confirm both months cover the same number of reporting days |
| The website had the largest increase | Website visits increased by a net 240 | Compare changes in sources and content |
| Paid campaign visits declined | Visits decreased by 68 | Add cost and conversion data before making a judgment |
This month's visits · Matches the original data
You can copy or save these sample materials. They illustrate the task instructions and result structure; they are not real customer data.
Match the method to the question
You do not need to start by asking for a “comprehensive analysis report.” Choose a suitable method and output first. That makes the calculation scope clearer and your follow-up questions more specific.
| What you want to know | How to analyze it | What the result should include |
|---|---|---|
| Totals and composition | Group and summarize by month, channel, or category. | Values for each group, the total, units, and whether the groups cover every record. |
| Where values increased or decreased | Compare two consistently defined periods, calculating both the absolute change and the percentage change. | The original values for both periods, the difference, and the denominator. Explain separately when the baseline is zero. |
| Whether a change is sustained | Examine consecutive periods to distinguish short-term fluctuations from a longer-term trend. | The full timeline, any missing periods, and changes in definitions. Do not select only the interval that supports a preferred conclusion. |
| Which records need investigation | Check extreme values, missing data, duplicates, and inconsistent combinations against business rules. | The location of each record, the rule it triggered, and what needs to be verified. Do not simply delete unusual records. |
For ratio metrics, recalculate the numerator and denominator. For example, the overall conversion rate is total conversions ÷ total visits; you generally cannot average the individual channel rates. Overall month-over-month change should also be calculated from the two period totals.

Turn the numbers into a useful discussion
In the example above, visits rose from 2,410 to 2,628, a net increase of 218. Website visits increased by 240, paid campaign visits fell by 68, and partner visits increased by 46. Those three changes still add up to 218. The website's increase is larger than the overall net increase because the decline in paid campaigns offset some of that growth. The numbers are consistent.
Download the app and bring your data
Define the input and the result you need, then refine the work based on the output.
Choose an Excel or CSV file. Identify the worksheet, time range, and what each row represents. Keep the original file and save the results separately.
Ask for the fields, record count, missing values, and duplicates. Confirm the file and definitions are correct before grouping or comparing data.
Summarize first, compare next, and then interpret. For example, total each channel's visits for both months before calculating the change. Ask separately for the results table, chart, and summary.
Spot-check original rows, totals, and ratios, and confirm the chart matches the table. If information is missing, add the data instead of asking AI to fill in a business conclusion.
More to know before you begin
This workflow starts with a file-based task in the desktop app and helps you define and carry out an analysis. Direct database connections, ongoing synchronization, enterprise metric systems, and online dashboards are separate capabilities. Support for local files does not imply support for those features.
Add the three channels for each month: 2,410 last month and 2,628 this month, a difference of 218. Then calculate 218 ÷ 2,410, which is about 9.0%. Each channel's change uses its own previous-month value as the denominator; do not simply average the percentages. The page uses sample data, and you can recalculate every value.
First, confirm that the file opens on your computer, identify the worksheet containing the data, and check the character encoding. Provide the file location, column headers, and field definitions. If the task cannot read or calculate the data, resolve that issue before generating a report. Do not substitute estimated numbers.
First, verify the sources, handling of missing values, aggregation definitions, and calculations. Correlated changes do not establish causation. Decisions involving costs, returns, or forecasts also require the relevant data and business judgment.
You can outline the questions, required fields, and report structure first. Do not present invented data as real findings. If you need an example, label the sample data clearly and replace it with verified data before making an actual decision.
You can start with a question in everyday language, but you still need to explain the fields and business definitions. Begin with a small, clear task—such as a monthly summary or a comparison of two channels—and gradually add conditions. Understanding the metrics matters more than using advanced terminology.
Start with a bar chart for category comparisons or a line chart for changes over consecutive time periods. Keep a table when readers need exact values. State the metric, units, and time period in the title. A chart does not replace the original data or the calculation definitions.
Open the Mockitt AI desktop app and describe the work you need to do.
Download Mockitt AI ↗