Profflow reads. It does not write an order, a price, a campaign, a payout or a setting back to any account it is connected to, and there is no plan on which that changes.
Platform / Integrations
Every figure comes
from somewhere you can check.
Profflow does not ask you to type your business into a settings page. It reads the accounts you already run — the store, the ad platforms, the payment processors — and each one is named here with what it reads, how often, and what it deliberately does not.
- 9connect today
- 0write to your accounts
- Nightlylast 7 days re-read
- 1source is enough to start
Read-only by default EU-hosted Disconnect at any time
01 / What a connection is for
Sources in. One model.
Figures you can open.
Nothing here is a dashboard of somebody else’s dashboard. Each source contributes the part of the arithmetic only it knows, and the model holds them together — which is why a missing source is visible rather than quietly assumed.
ShopifyMeta AdsGoogle AdsStripePayPalKlarna
KlaviyoOne modelEvery source in one set of figures
Net profit, measuredContribution per productReturn per channelVAT owed, all quarterCash, six months out02 / The rules every connection follows
The same six rules,
whatever you connect.
Each connection asks for the narrowest permission that produces the number. No customer list is copied for its own sake, and nothing is read that does not appear in an answer you can open.
A charge read from an account is measured and is labelled so. Anything modelled is labelled too, in words, on the screen it appears on.
A source you do not connect is simply not in the figure, and the screen says which sources are in it. There is no bundle, and no connection is a prerequisite for another.
Disconnecting stops the reading immediately. What was already computed stays, so last quarter does not vanish because you changed payment processor.
European infrastructure, GDPR-aware handling, and a written record of what is held and for how long. The security page has the detail.
03 / Every source, in detail
What each one reads.
Status is what it says: a source waiting on somebody else’s approval says so rather than carrying a badge that sends you to a screen that refuses you.
Commerce
Where the orders are. Shopify is enough to run the whole model; everything below it adds to the figure rather than being required by it.
ShopifyLiveReadsOrders with their lines, discounts, shipping and destination country; refunds against the order they reverse; products and variants; tax treatment per order.
Cost of goods is read from Shopify where you keep it there. Where you do not, it is entered in Profflow with the date it takes effect.
Nightly, last 7 days re-readHow it is used
Advertising
Spend belongs in the same period as the revenue it bought. Until it is, a good day and an expensive day look identical.
Meta AdsLive
ReadsDaily spend by account and campaign, converted to your currency at the day’s rate.
Spend as the platform billed it. Profflow does not take Meta’s attributed revenue as fact; it measures the channel against contribution.
Nightly, last 7 days
Google AdsLive
ReadsCampaign cost by day, in the account’s currency, converted at the day’s rate.
The same treatment as Meta: cost is measured, credit for revenue is not assumed.
Nightly, last 7 days
TikTok AdsWaiting on approval
ReadsDaily spend by advertiser account, converted at the day’s rate.
Built and waiting on TikTok’s own developer approval. Until then TikTok spend can be entered as a cost so the month is still complete.
Nightly once open
Payments
A typed percentage is a guess. The processors report what they actually charged, transaction by transaction.
StripeLive
ReadsCharges and refunds with the fee taken on each, disputes, and payout timing.
Connected through Stripe’s own authorisation. Read-only: no charge, refund or payout can be created from Profflow.
NightlyHow it is used
PayPalLive
ReadsTransactions with their actual fees, refunds and the economics of a reversed sale.
PayPal fees vary by country, currency and transaction type, which is exactly why they are read rather than assumed.
NightlyHow it is used
KlarnaLive
ReadsSettlements with the fee per order, returns, and the delay before the money lands.
Buy-now-pay-later settles late. The cash forecast uses the real delay rather than treating the sale as paid on the day.
NightlyHow it is used
Lifecycle
Email is a channel with a cost, and should be judged like one.
KlaviyoLiveReadsCampaign and flow revenue day by day, with conversions, beside the plan cost you enter.
This is attributed revenue — Klaviyo counts an order the customer might have placed anyway — so it sits beside ad spend and is never added to profit as though it were extra.
Nightly, last 7 daysHow it is used
Your figures elsewhere
The model is not a place you have to work. It is a place the figures are correct.
Google SheetsLive
ReadsOrder-level profit, revenue, ad spend and costs, pulled into a sheet with a script and a key.
A read key, rate-limited. The sheet reads Profflow; Profflow does not read the sheet.
Refreshed each morning
API and exportsLive
ReadsThe same figures the screens use, as CSV or through an API key you create and can revoke.
Useful for an accountant, a warehouse or a board pack. Keys are per-account and can be revoked without disconnecting anything else.
On request
04 / From authorising to trusting the figure
Five steps, and four of them
happen without you.
- 01AuthoriseThrough the source’s own screen — Shopify, Stripe, Meta — or with a read key you create. Profflow never asks for a password to an account it is connecting to.
- 02BackfillHistory is read as far back as the source will give it, so the first screen you see has a past as well as a present.
- 03ResolveEach order is matched to the cost in force on its date, the VAT rate of its destination, and the fee the processor charged on it.
- 04RefreshEvery night the last seven days are read again, so a late refund or a corrected charge lands in the month it belongs to.
- 05CheckEvery figure opens to the rows beneath it, and the reconciliation screen compares what the model says arrived with what the payout actually was.
05 / What Profflow will not do
The part most
integration pages leave out.
No order, price, campaign, payout or product is ever changed by Profflow. The permission is not requested, so it cannot be misused later.
Your figures are yours. They are not aggregated into a benchmark product, sold to a third party, or used to train anything.
Connect one source or six. What is not connected is not in the figure, and the screen says which sources are.
If a cost is missing, the gap is shown rather than filled with an assumption. A margin with a guess inside it is worse than no margin at all.
Questions
Before you connect anything.
- What is the minimum to get a real figure?
- One commerce source. Shopify alone gives net profit after goods, refunds and VAT; adding the ad accounts and the payment processors replaces the last estimates with measured charges.
- How long does a connection take?
- Minutes. Authorise, and the current period appears immediately while history backfills behind it.
- What if I use a payment processor you do not read?
- Enter the fee as a cost and the month stays complete — it is simply entered rather than measured, and the screen says so. Tell us which one and it goes on the list.
- Does connecting more sources cost more?
- No. The plan is the plan; sources are not priced individually.
- What happens when I disconnect?
- Reading stops immediately. Figures already computed stay, so your history does not disappear because you changed a tool.
- Where is the data held?
- On European infrastructure, with GDPR-aware handling and a record of what is held and for how long.
Start with one source.
Shopify alone produces a real net profit figure. Everything else replaces an estimate with a measured charge, and none of it is required.