Write to me

Numbers You Can Bet the Business On

Imagine your next meeting starting with numbers everyone already trusts.

Did Shopify, Meta and your accounts disagree on what your webshop sold last month?

Do your reports, AI answers and board decks already use the same checked, current numbers? Then you don't need this.

Write to me

A meeting room with a long oak table, green chairs and closed laptops, a window over rooftops and spires

Your own AI, answering from numbers you trust.

What were net sales in September's third week, and how did they change?

1.452.000 DKK

Down 21% from 1.843.000 DKK the week before, when the autumn collection launched with the wool coat selling most. Sales fell back after the launch. September's net sales were 6.420.000 DKK.

From your metrics book: Net sales by week and for September.

  • Claude or ChatGPT

    Use the AI you already have, with answers from the same metrics book.

  • Follow your next question

    Go from September's total to a single week, then ask which styles or countries changed.

  • Ask in your own words

    You can ask as you would ask a colleague, without learning how to pull the numbers.

Ask your AI: What were September's net sales?
  1. Through Meta7.340.000 DKKIt counts sales after an ad click or view, including orders later returned
  2. Through the Shopify connector9.170.000 DKKIt counts orders placed, excluding VAT, before returns
  3. Through the accounts6.420.000 DKKIt reads invoiced sales after returns, a month late
  4. Through your metrics book6.420.000 DKKIt uses one written definition, with every difference explained

Your AI needs more than the Shopify connector.

Connecting your AI to Shopify doesn't settle how to count returns or bring in costs from other systems. More connectors still leave your AI choosing between different answers.

I put the definitions in your metrics book, so you and your team can ask the same question and act on the same answer, checked against the accounts.

Your own view of net sales
MarketingNet sales by channel
FinanceNet sales by month, ending in September
MerchandisingNet sales by style, last week, share of net sales
Wool coat12%
Knit vest10%
Merino sweater6%

All three read Net sales from the metrics book, so they agree on September's 6.420.000 DKK.

Keep the way you like to work.

If your merchandiser prefers a spreadsheet, that can stay. You shouldn't have to teach the whole team a new tool before they can use the numbers.

They can ask for the view they need and carry on with the job.

Your weekly meeting notes prepare themselves.

Your weekly meeting notes, ready before you sit down

Net sales, last week

1.480.000 DKK

+6% on the same week last year

Return rate, orders past their return window

31%

+2 percentage points

Net sales by style, last week

  1. Wool coat
  2. Knit vest
  3. Merino sweater

Weeks of stock left

14 weeks left, and falling

September's net sales are 6.420.000 DKK. Your board report reads the same number.

  • Ready when you sit down

    Your numbers are up to date within minutes, with notes ready before your weekly meeting.

  • Spot what needs attention

    See where sales, returns or stock need a decision before the week gets away from you.

  • Fewer reports to read

    I keep the reports that help you make a decision and cut the ones nobody uses.

Net sales by week, July to September

FounderThe autumn collection launched. The wool coat sold most.

MarketingCheck which channels explain the drop.

This yearLast year

These weekly net sales agree with the accounts, as does September's 6.420.000 DKK.

Remember what changed when sales moved.

A good week is easier to repeat when you know what you did. Your team can leave a note beside the week it happened.

When you look back, you can judge the next campaign with that history in front of you.

A long oak table with closed laptops and mugs behind a glass wall onto a large, bright warehouse

Your board deck explains what changed.

September board deck

Sales rose with the autumn collection launch.

September net sales

6.420.000 DKK

+11% on last September

  1. Week 1
  2. Week 2
  3. Week 3
  4. Week 4

SeptemberLast September

  1. 1

    The autumn collection launch lifted week 2 sales, with the wool coat selling most.

  2. 2

    Sales fell in week 3 after the launch, but stayed above the same week last year.

Returns rose as launch orders came back.

September return rate, orders past their return window

31%

+2 percentage points on last September

  1. Week 1
  2. Week 2
  3. Week 3
  4. Week 4

SeptemberLast September

  1. 1

    Returns rose in week 3 as orders from the autumn collection launch came back.

Stock cover fell to 14 weeks.

Weeks of stock left

14

Down from 19 at the start of September

  1. Week 1
  2. Week 2
  3. Week 3
  4. Week 4
  1. 1

    Stock cover fell in week 2 because the launch lifted wool coat sales.

  2. 2

    Stock cover fell again to 14 weeks as sales continued after the launch.

Your AI writes the reasons from the metrics book. The deck goes to your board and investors automatically in early October.

  • Go beyond the total

    Explain the month by week, style or country, with the detail ready when your board asks.

  • Reasons beside the numbers

    Your AI writes why each number moved, using the same metrics book that supplies the figures.

  • Stop rebuilding each month

    Your monthly deck prepares and sends itself, so repeated reporting no longer takes your time.

One metrics book behind every answer.

September net sales, 6.420.000 DKK, wherever you read it

Where it comes from

  • SalesShopify
  • MarketingMetaGoogleTikTokKlaviyo
  • Moneye-conomic
  • OperationsWarehouse PDFsCarrier invoicesDuty

Your metrics book

Your numbers

In your name. Each number defined once, checked every morning.

  1. Net saleswithin minutes
  2. Gross margindaily
  3. Return ratedaily
  4. Cost per orderdaily
  5. Marketing costdaily
  6. New customerswithin minutes
  7. Repeat rateweekly
  8. Stock coverdaily

Where you read it

  • Your weekly meetingready before you sit down
  • Your own AIask Claude or ChatGPT what changed
  • The board reportsent to your board and investors each month
  • Merchandisingsee what sells before the next order
  • Your shared reportsthe ones your team uses to make decisions
  • Keep the systems you use

    Shopify, your accounts and your warehouse stay in place while their numbers come together.

  • Trust tomorrow's numbers too

    Checks run on their own each morning, so a broken update doesn't quietly become your next decision.

  • The software is included

    Your price includes hosting, the warehouse and every connection needed to keep your numbers up to date.

What your numbers mean, and who owns them
NumberWhat it meansOwner
Net salesWhat customers paid, excluding VAT, less what they sent back.Finance
Return rateItems sent back, divided by items sold.Operations
Cost per orderPicking, packing, freight, duty, returns and payment fees for one order.Operations
Marketing costEvery DKK spent on ads and agencies. Agency fees count the month they are billed.Marketing

Stop being the person everyone has to ask.

You used to hold the whole business in your head. Now each new colleague needs to learn what you already know before they can make a call.

I get that knowledge into the book, so they can find an answer without waiting for you.

Get 38 written definitions you can change to fit your shop and export for free. Start my free metrics book.

What each order earns once returns and every cost are taken off is in Steer by Profit, Not Sales.

A loading dock with its door open, shrink-wrapped pallets of cartons and a pallet jack on the floor

At Miinto, I started with what each number means

When I took over the numbers at Miinto, they were a mess.

Nothing was written down. The same number came in many versions. Reports were slow, and they broke a lot. It took too many people to keep them going.

Worst of all, nobody agreed on which number was right. So we made wrong calls, even with lots of data.

So we started with the basics. We agreed on what to steer by, and we wrote down what each number means.

Then we built from the orders up. First the orders and what was in them. Then the products and the customers. Then payments, the books and the cost of ads. Everything in a webshop hangs on its orders, so I always start there.

Today each number has one meaning, in one metrics book. Most questions go to Claude. It reads the same meanings, and it checks what you may see. So two people who ask get the same answer.

We built a tool on it called Power House. Each team sees its most important numbers next to its goals. We use it in our meetings, and we write notes on what looks wrong and what we will do.

Not all of it was easy. The hardest part was people. If you have read numbers one way for years, a new way feels wrong. That took longer than I hoped.

Some things are still hard. Does a sale count on the day of the order, or come off when the item comes back? There is no right answer, so you pick one and write it down. Systems change, and the setup can break when they do. It needs more care than most people think. And Claude still picks the wrong number now and then, so we check the answers that matter.

That care is what most setups never get. In yours, it is my job.

Every tool reads the same metrics book, so a new tool has the right numbers from day one. Nobody builds them again. This is how we build Tools as a Service.

This is the ground everything else stands on. Without it, we could never have built Steer by Profit, Not Sales.

  • 20.000

    DKK a month, from the day it's live

  • 0 DKK

    extra for software. It's included in the price, and your AI uses your existing subscription.

  • Week 4

    when it's live, by a definition we agree before we start

Your data is stored in Europe, in a setup only your company uses.

  • Hosted in Europe

    Your data, your metrics book and your tools all run on servers inside the EU.

  • Yours alone

    No other client's data sits next to yours, and the whole setup runs in your company's name.

  • You decide who sees what

    Each person, and each person's AI, reaches only what their job needs.

  • Built for a GDPR or security audit

    We sign a data processing agreement before we start, and the setup is built the way auditors expect, in the EU, apart from other clients, with access set person by person.

Two exceptions, both services your team already uses. When you ask Claude or ChatGPT, the question and its answer may be handled outside the EU. And Meta and Google receive what each order earns, so they can bid on it.

Tell me which number starts the argument in your meetings.

Write a few lines about the business and what you can't answer today. I'll reply myself, and if it's a fit we book a call.

Write to meBook a 15-minute intro call