Example questions
Not written for this page. Each answer below is the verbatim output of the same pipeline that answers you in the app, in Telegram, or through the MCP server inside your coding agent — down to the wording and the tables.
They come from Lumen — Sleep Tracker, the demo account we keep for testing and screenshots. The app is fictional and its data synthetic; the analysis over it is not. We use it because a customer's revenue is theirs, not ours to publish.
3 of the 10 tell the account something it would rather not hear — cohorts monetising worse, flat DAU, a comparison refused as misleading. A page where the tool always agrees with you says nothing about what it does when the data disagrees.
What's my MRR right now?
Your current MRR is $9,956.58 as of 2026-09-02 (today, based on data synced ~11:54 UTC).
For context, it’s been climbing steadily over the past week — from $9,461.98 on 8/26 to $9,956.58 today, driven mostly by new subscriber MRR outpacing churn each day (e.g. today added $39.98 in new MRR with $0 churned).
Split my MRR change last month into new, expansion and churned
In August, MRR grew from $7,333.81 (end of July) to $9,836.68 (end of August), a net gain of +$2,502.87 (from tool results directly).
Decomposition of the change (August flows):
Essentially all of August’s MRR growth came from new payers, with churn offsetting a portion of that. There was no expansion revenue (upgrades/price increases) this month. Comparing to July: new MRR was $3,407.13 and churned MRR was $249.78 — so new-MRR was actually higher in July than August, while churned MRR roughly doubled from $249.78 to $459.56.
Compare this month's revenue to last month
Month-to-date, September revenue is $879.42 (Sep 1–2 only) vs August’s full-month $13,930.71 — but that’s not a meaningful comparison since September has just started.
A better apples-to-apples comparison: August vs July (both full months) — revenue was essentially flat: $13,930.71 in August vs $13,912.58 in July, a delta of +$18.13 (+0.13%).
So if you’re asking whether revenue is trending up or down, the honest answer is: revenue has been flat month-over-month (July→August), and it’s too early to say anything about September with only 2 days of data. I’d recommend checking back after September has more days elapsed for a real comparison.
Data notes: No caveats or partial-data flags were returned by the tool for either comparison — both are marked “complete.” The September figure itself only reflects 2 calendar days, not a full month, so treat it as directional/in-progress rather than a finished period.
What's my trial-to-paid conversion rate?
Your trial-to-paid conversion rate is running around 43–45% for cohorts that have had time to fully convert.
Looking at daily trial-start cohorts from Aug 1 – Aug 26, 2026 (the ones old enough to finish converting, since trials typically take ~7 days to resolve):
Important caveat: cohorts from Aug 27 – Sep 2 show 0% conversion — this is not a real drop, it’s right-censoring. Trials started in the last ~7 days haven’t had time to convert yet, so those rows will fill in as time passes. Don’t read them as churn or a funnel break.
Data notes: Trial conversion is denominated by trial-start cohort, not conversion day. The most recent ~7 days of cohorts (Aug 27–Sep 2) are still in-flight/censored and should not be treated as 0% conversion — they will rise as trials resolve.
What's my customer LTV?
Customer LTV depends on cohort age — it grows over time as more revenue accrues, and your oldest, most-matured cohort (May 2026) sits at $44.72/user after 4 months, against a blended CAC of $24.25, for an LTV/CAC ratio of 1.84.
Cohort detail (cumulative net revenue per user, by first-payment month):
| Cohort | Users | Latest LTV/user | Months observed | LTV/CAC |
|---|---|---|---|---|
| 2026-05 | 200 | $44.72 | 5 (through month 4) | 1.84 |
| 2026-06 | 308 | $41.05 | 4 (through month 3) | 1.69 |
| 2026-07 | 406 | $34.09 | 3 (through month 2) | 1.41 |
| 2026-08 | 343 | $26.81 | 2 (through month 1) | 1.11 |
| 2026-09 | 16 | $28.74 | 1 (month 0 only) | 1.19 |
CAC is flat at $24.25 across all cohorts (same blended figure applied). The younger cohorts show lower LTV simply because they haven’t had time to accrue revenue yet — not because they’re performing worse. Comparing at the same month-index (e.g., month 0: $24.24 → $25.31 → $24.74 → $23.05 → $28.74, or month 1: $31.08 → $32.00 → $31.24 → $26.81), the trajectories look broadly consistent, though the August cohort’s month-1 value ($26.81) is a bit softer than earlier cohorts’ ($31–32) — worth watching as it matures.
Bottom line: mature-cohort LTV ($45 at 4 months, still climbing) comfortably exceeds blended CAC ($24), giving a healthy >1.8x ratio for the oldest cohort, but recent cohorts are still too young to know their eventual LTV — their current ratios (~1.1–1.4x) will rise as revenue keeps accruing.
Are newer cohorts monetizing better or worse than older ones?
Newer cohorts are monetizing worse — LTV per user at comparable time-since-cohort points is trending down, and the LTV/CAC ratio is shrinking each month.
LTV trajectory by first-payment cohort (per-user cumulative net revenue):
Month-0 LTV is fairly flat (~$23–28 across all cohorts), but each successive cohort’s revenue growth after month 0 is weaker — e.g. May grew ~85% from month 0 to its latest point, June ~62%, July ~38%, August only ~16% (one month in). This points to either declining post-trial retention/expansion or slower upsell/renewal behavior in newer users, not an acquisition-quality (CAC) problem, since CAC is held flat at $24.25 across all cohorts in this view.
Trial→paid conversion (a leading indicator) confirms the softening: daily cohort conversion rates in June averaged roughly mid-40%s to 60%s (e.g. 6/12 at 61%, 6/21 at 64%, 6/30 at 69%), July cohorts clustered more in the 35–55% range, and August cohorts skew further down (many in the 30–45% range, e.g. 8/13 at 31%, 8/20 at 33%, 8/26 at 33%). Late-August/September cohorts show 0% converted, but that’s expected right-censoring — those trials haven’t had the ~7 days needed to convert yet, not a real drop to zero.
Likely explanation (moderate confidence): the combination of flat month-0 revenue, decelerating cumulative LTV growth, and a downward drift in trial-conversion rates suggests newer user cohorts are lower-intent or lower-quality on average (possibly tied to changing acquisition mix/campaigns), rather than a pricing or product change — but I don’t have campaign-mix-by-cohort-month data in this pull to confirm the cause. I’d recommend checking whether ad spend/campaign mix shifted materially into July–August before concluding it’s an acquisition-quality issue.
Data notes: All figures above came back “complete” with no caveats. However, the September 2026 LTV cohort (16 users) and the trailing trial-cohorts from ~2026-08-27 onward have not had time to mature/convert (right-censored) — their 0% conversion and single-point LTV are not evidence of decline, just insufficient elapsed time. Treat July/August LTV trajectories as directional, not final, since later months of their curves haven’t accrued yet either.
What does my install-to-paid funnel look like?
Here’s the install-to-paid funnel snapshot for 2026-08-01 to 2026-09-02 (as of 2026-09-02):
| Step | Users | Conv. from previous | Conv. from first |
|---|---|---|---|
| App opened | 5,757 | – | – |
| Paywall viewed | 4,310 | 74.9% | 74.9% |
| Trial started | 2,949 | 68.4% | 51.2% |
| Subscribed | 1,273 | 43.2% | 22.1% |
Overall, 22.1% of users who opened the app end up as paying subscribers. The biggest drop-off is post-trial: only 43.2% of trial starters convert to paid, while paywall exposure (74.9%) and trial start (68.4% of paywall viewers) both hold up reasonably well.
Likely next step: the trial→paid step (56.8% loss) is where the funnel leaks most — worth digging into trial length, onboarding/engagement during the trial, or reminder/paywall timing near trial expiry. This is a snapshot, not causal, so I’d treat it as a place to investigate rather than a diagnosed cause.
Data notes: This funnel counts only subscriptions linked to an app user (per-user step intersection). If any store purchases aren’t SDK-linked, they’d still show up fully in get_subscription_stats but not here — so the true paid count could be modestly understated. This particular result was marked complete with no caveats, but keep that linkage caveat in mind when comparing to raw subscription counts.
Is my DAU growing week over week?
No — DAU has been roughly flat, not growing, over the last four weeks.
Weekly average active users (Tue–Mon blocks, Aug 5 – Sep 2, 2026):
DAU has oscillated in a tight ~950–1,080 band all month with no clear upward trend — if anything, the most recent full week (Aug 26–Sep 1) is slightly below the Aug 12–18 peak, a dip of roughly 4%. New user counts are similarly steady (~40–59/day) rather than accelerating, which lines up with flat DAU — the app isn’t pulling in enough incremental new users to grow the active base, and retention/usage frequency also isn’t compensating.
Confidence: high on the numbers themselves (complete data, no caveats); the “why” (acquisition vs. retention) would need a look at get_retention or campaign trends if you want to dig deeper.
Why did my MRR grow last month?
MRR grew from $7,418.73 (Aug 1) to $9,836.68 (Aug 31) — up $2,417.95 (+32.6%) — driven almost entirely by new-payer acquisition, with churn being a modest, steady drag.
Breaking down the daily flows summed over August:
Net: $2,962.42 new − $459.56 churned ≈ $2,502.86, which reconciles closely with the observed $2,417.95 rise (small day-to-day rounding/timing at the boundary).
Takeaway: growth last month was a pure new-customer story — no expansion revenue (no upsells/price hikes) and churn stayed proportionally small and stable rather than the cause of any slowdown. If you want to accelerate MRR further, the lever isn’t retention (already fairly stable) — it’s likely acquisition volume, since expansion is completely untapped.
Which campaign is responsible for yesterday's spend increase?
Total ad spend rose $22.65 from Aug 31 → Sep 1 ($240.17 → $262.82). The increase was driven almost entirely by Sleep Better — iOS — US (Meta), which jumped from $89.90 to $123.64 (+$33.75). That was partially offset by a drop in Search — Sleep Tracker — US (Google Ads), down $14.25 ($80.72 → $66.47), while Insomnia Interests — Broad (Meta) added a smaller +$3.16.
Net effect: Sleep Better — iOS — US alone more than accounts for the total increase; the other two roughly cancel each other out.
Pulse cannot pause a campaign, move a budget or change a price, and asking produces a refusal plus an offer to analyse the decision instead — never a claim that it did something.
Competitor numbers, forecasts of future values, app-store keyword research: a short "that's outside the data I can see", not an invented answer.
Every number is checked against the query results that produced it. If a check fails the answer is not sent: the model is asked to redo it, and failing that you get the raw query results instead of prose. You are never shown a figure that came from the model rather than from your data.
Open Pulsefrom a coding agenton Telegram
New account with nothing in it yet? Start with the Quickstart and connect a source — the answers above need revenue and ad data behind them.
