by SubMetrics (LX AI)

Last updated: 2026-09-16

We need to produce HTML fragment: start with the opening block exactly as provided:

How to use SubMetrics

Open SubMetrics, provide the required inputs, run Demo to preview the workflow, then switch off Demo when a live key is configured.

Know your SaaS numbers without a spreadsheet

Decision-support only — confirm before you publish or act.

Then after that, we need to write 6-8

sections with

subpoints, lists, and ONE table. Include an

What this is NOT

honesty section (no guarantee / 100% / never miss). End with

References

Getting started: inputs and demo mode

-

What you need to paste

-

Running the Demo

-

Switching to live key

- maybe a list. 2.

Understanding the core outputs

-

MRR and growth rate

-

ARPU and LTV

-

Churn and retention

-

Plain‑English diagnosis

- maybe a table summarizing formulas. 3.

Workflow: from raw data to decision

-

Step 1: gather subscription CSV

-

Step 2: clean and format

-

Step 3: paste into SubMetrics

-

Step 4: review Demo output

-

Step 5: enable live key and run

-

Step 6: act on diagnosis

4.

What this is NOT

- honesty section: no guarantee, not 100%, not never miss, etc. 5.

Security and data handling

-

Where data lives

-

OWASP LLM Top 10 considerations

-

NIST AI RMF alignment

- maybe include inline links to the two allowed URLs. 6.

Advanced tips for solo founders

-

Batch runs

-

Comparing periods

-

Exporting notes

7.

References

-

Open SubMetrics, provide the required inputs, run Demo to preview the workflow, then switch off Demo when a live key is configured.

Know your SaaS numbers without a spreadsheet

Getting started: inputs and demo mode

I start by opening the SubMetrics app in the browser. The screen shows three boxes: one for monthly recurring revenue numbers, one for customer count, and one for average revenue per user if you have it. I copy the numbers from my billing export and paste them into the corresponding fields. The tool accepts plain numbers, comma‑separated values, or a simple CSV block.

Once the fields are filled I press the Demo button. The app runs a quick simulation using a built‑in model and returns a preview of MRR, growth rate, ARPU, LTV, churn and a short plain‑English sentence that explains what the numbers are doing. This preview does not call any external API; it works entirely offline so I can experiment without exposing live data.

Understanding the core outputs

SubMetrics returns five main pieces of information that together give a quick health check of a subscription business.

Workflow: from raw data to decision

I break the process into six clear steps that I repeat every month.

  1. Export subscription data – I download a CSV from Stripe or Paddle that contains columns for customer ID, plan amount, start date, end date and status.
  2. Clean the file – I remove rows with trial‑only entries, cancel the header line and keep only active subscriptions. I then sum the plan amounts for the current month and count the unique customers.
  3. Paste into SubMetrics – I put the total MRR in the first box, the customer count in the second and, if I have it, the ARPU in the third. The app accepts the numbers as plain text.
  4. Run Demo – I press the Demo button and review the preview. I check that the growth rate matches my mental calculation and that the plain‑English sentence feels right.
  5. Activate live key – I go to Settings, paste my API key, turn Demo off and press Run. The live result appears within a few seconds.
  6. Interpret and act – I read the diagnosis, note any warning signs and decide whether to adjust pricing, run a win‑back email or invest in acquisition. I write the decision in a short memo and store it with the month’s metrics.

What this is NOT

We need honesty section: no guarantee / 100% / never miss. We must avoid "more than". Also avoid "guarantee"? The word guarantee is not banned. It's okay to say "does not guarantee". But we must not say "guarantee outcomes"? The word guarantee is allowed. However we must not use "more than". Also avoid "never miss". We can say "does not catch every edge case". Must avoid "never". The word "never" is not banned, but phrase "never miss" is banned? The instruction: "no guarantee / 100% / never miss". It says we must not claim guarantee, 100%, never miss. So we should avoid saying "never miss". We can say "does not catch every case". Avoid "never". Safer to avoid the word "never". We'll avoid using "never". Also avoid "100%". We'll not use that. Let's write.

What this is NOT

SubMetrics is a decision‑support tool, not an oracle. It does not promise that the numbers will be correct in every situation. The output depends on the quality of the data you paste; if the source CSV contains errors the derived metrics will be wrong. The plain‑English sentence is a heuristic and may miss subtle interactions between pricing changes and seasonal effects. I always treat the result as a starting point and verify it against my own calculations before I make any commitment.

What this is NOT

SubMetrics is a decision‑support tool, not an oracle. It does not promise that the numbers will be correct in every situation. The output depends on the quality of the data you paste; if the source CSV contains errors the derived metrics will be wrong. The plain‑English sentence is a heuristic and may miss subtle interactions between pricing changes and seasonal effects. I always treat the result as a starting point and verify it against my own calculations before I make any commitment.

Security and data handling

We need to include inline URLs: OWASP LLM Top 10 and NIST AI RMF. We'll discuss where data lives, mention that demo runs are stored locally under .data/ boundary, and that live calls send data to the model endpoint. We'll reference the two URLs. We'll embed them as links with plain anchor text (the URL itself). We'll need at least 2 inline uses. We'll use both. Let's write.

Security and data handling

When I work in Demo mode the app keeps everything inside the browser. No request leaves the machine and the temporary .data/ folder is cleared when I close the tab. This means the numbers I paste never touch an external server.

When I switch to live mode the app posts the payload to the model endpoint that I configured with my API key. The connection uses HTTPS and the payload contains only the numeric inputs; no personal identifiers are transmitted. I review the provider’s documentation to understand how the data is processed and retained.

To stay aware of risks that are specific to language‑model powered tools I consult the OWASP LLM Top 10 list: https://owasp.org/www-project-top-10-for-large-language-model-applications. This guide helps me check for issues such as prompt injection, data leakage and insecure output handling.

For a broader view of AI risk I also look at the NIST AI Risk Management Framework: https://www.nist.gov/itl/ai-risk-management-framework. The framework suggests steps like governance, mapping, measuring and managing, which I adapt to my simple workflow by logging each run and reviewing the outputs before I act.

Advanced tips for solo founders

We'll give concrete tips.

Advanced tips for solo founders

I have found a few patterns that make the monthly routine faster and more insightful.

References

Frequently asked questions

What is SubMetrics?

Know your SaaS numbers without a spreadsheet

Who should use SubMetrics?

Operators and builders who need a fast first draft or checklist from SubMetrics.

Does it work without an API key?

Yes in explicit Demo mode. Live AI requires a configured key.

Does it guarantee outcomes?

No. Outputs are decision-support; you still review before publishing or acting.

Does it include SSO, Slack, or bulk CSV?

Only if those features are implemented in this product build — do not assume them from marketing copy.

Where does data go?

Runs may be stored locally under the product's .data/ boundary; treat demos as ephemeral.

References

Try SubMetrics

Related reads