# How to Create a Sales Forecast for an Ecommerce Brand From Your Own Order History

E-commerce sales forecasting for Shopify brands: read next quarter in a revenue forecast built from your own orders and seasonality, check it against actuals, then set revenue targets and a stock budget.

- Applies to: Brands for which Luca's team has switched forecasting on
- Status: Live
- Verified: 2026-11-07
- Canonical page: https://ask-luca.com/product-docs/use-cases/sales-forecasting-for-ecommerce

E-commerce sales forecasting starts with a question every Shopify founder asks in November: what will next quarter bring once peak is over? You need one number for the team's revenue targets and another for the stock you can afford to buy. Luca AI builds that forecast from your own Shopify orders, allows for seasonality and holidays, and checks it against actual sales every night. This is ecommerce revenue forecasting from your store's orders, not B2B sales-pipeline forecasting from CRM deals.

## At a glance {#at-a-glance}

| Question | Answer |
|---|---|
| What you forecast | Your brand's revenue, daily and up to a year ahead, so a full quarter is covered |
| What it is built from | Your own Shopify order history, with seasonality and holidays allowed for |
| Where you read it | **Numbers**, then the **Forecast** tab, or by asking in chat |
| How far to trust it | The forecast is checked against actual sales every night and corrected |
| What you compare it with | The same weeks last year, plotted on the same chart |
| What you decide | Your team's revenue targets and the stock and cash you commit to |
| What it is not | B2B pipeline forecasting from CRM deals and stages |
| Who switches it on | Luca's team, per client |

## Where it fits {#where-it-fits}

### What it's for {#what-its-for}

E-commerce sales forecasting turns your order history into a number for the weeks ahead. You use it to set next quarter's revenue targets and to size the stock and cash behind them. The aim is a number you can explain, not a hopeful one.

### How it helps {#how-it-helps}

- **Targets start from your own seasonality.** A January after peak is judged against your past Januaries, not against December.
- **You see last year on the same chart.** The gap between the forecast and last year is the growth you are assuming.
- **You know how far to trust it.** A forecast checked against actuals every night shows whether it has been running high or low.
- **Stock and cash follow the same number.** The stock budget comes from the forecast your team is chasing, so the two plans agree.

### What happens without it {#without-it}

Without a forecast, next quarter's targets are a guess dressed as a plan. Set them high after a strong peak and you buy stock that sits through a quiet January, with the cash locked inside it. Set them low and you run short of bestsellers when spring demand returns.

### Why buyers care {#why-buyers-care}

This matters most to founders of Shopify brands with a sharp seasonal curve, such as apparel, gifting and beauty. It bites hardest in November and December, when peak trading makes every week look like the new normal just as next quarter's plans are due.

## The Luca AI approach to e-commerce sales forecasting {#approach}

{/* 5P */}

Picture the founder of a Shopify apparel brand in mid November. Knitwear and outerwear are selling fast, and the team wants next quarter's targets before the Christmas break. She also has to approve the spring stock budget, knowing January is always slow. She needs one figure for both decisions.

The usual answer is a spreadsheet: last year's first quarter plus a growth rate. The growth rate tends to come from the strongest recent weeks, which are peak weeks. That can lift January targets well above what January has ever done, and the stock bought against them can sit until spring.

The logic of the situation points somewhere simpler. Your order history already holds the shape of your year: the peak, the January dip and the spring return. A forecast built from that history keeps the dip in view. A forecast checked against what then happened tells you how much weight it can bear.

Luca AI builds that forecast from your Shopify orders. It forecasts daily revenue up to a year ahead and allows for seasonality and holidays such as Black Friday. It checks every forecast against your actual sales each night and corrects it. You read the result on the **Forecast** tab, or ask about it in plain English in chat.

You can check the starting point yourself. The **Revenue Forecast** card plots actual revenue, the same weeks last year and the shaded forecast on one weekly chart. That combination of your own history, seasonality and a nightly check against actuals is covered in How Luca AI compares for e-commerce sales forecasting, below.

## Start from your own order history {#order-history}

A sales forecast for an ecommerce brand is only as good as the orders behind it. Your store is the record of what sold, when and at what price. That record already shows how your peak builds, how deep your January dip goes and when spring demand comes back.

Luca imports your full Shopify history when you connect the store. Every source lands in one private warehouse for your brand and refreshes once a day, overnight. The forecast, your dashboards and chat all draw on that same warehouse, so the figure in a chat answer matches the one on the **Forecast** tab. If you have not connected yet, [setting up Luca](https://ask-luca.com/product-docs/getting-started/set-up-luca) walks through the store connection.

Before you trust any forecast, check three things in your own history:

1. **At least one full year of orders.** A quarterly forecast needs last year's version of the same quarter to compare with.
2. **One-off events you know about.** A viral product, a stock-out or a launch can distort a single month.
3. **Changes the history cannot know.** A new price, a new category or a planned promotion all change next quarter.

Luca only reads your store data and never changes your store. The [read-only access page](https://ask-luca.com/product-docs/platform/read-only-access) explains what it can and cannot touch.

## How ecommerce sales forecasts are built {#how-forecasts-are-built}

Most brands set next quarter's numbers with one of four methods, whether they name them or not. Each one handles a January after peak differently. The method behind your number matters more than the spreadsheet it sits in.

### Last year plus a growth guess {#last-year-plus-growth}

This is the most common method. You take last year's quarter and add a growth rate. Statisticians call the first half a [seasonal naive forecast](https://otexts.com/fpp3/toolbox.html): each forecast equals the last value from the same season, such as the same month last year.

It catches your seasonal shape, because last January was also slow. It misses anything that has changed since, and it depends entirely on the growth rate you pick. Pick that rate from peak weeks and the whole quarter inflates.

### Recent run rate {#run-rate}

Here you take recent weekly sales and carry them forward. The textbook version, the naive method, sets every forecast to the last observed value. A drift version adds the average change seen in the history.

It catches momentum, which is why it feels convincing in November. For a January target it is the riskiest method, because the most recent weeks are your biggest of the year. It carries peak into the dip.

### Average of past months {#average-of-past-months}

Some teams average the last six or twelve months. The average smooths out one-off spikes.

It misses seasonality almost entirely. Averaged over a year, January looks like an ordinary month, so the target is too high in January and too low in spring.

### A seasonal model from order history {#seasonal-model}

Models such as [Holt-Winters exponential smoothing](https://otexts.com/fpp3/exponential-smoothing.html) estimate a level, a trend and a seasonal pattern from the history itself. Holidays can be added as known events.

This catches the dip after peak and the trend underneath it at the same time. It misses what is not in the data, such as a price change you plan for February. That is why a forecast should list the assumptions it was built on.

### What this means for next quarter's targets {#what-this-means}

For a slow January after peak, the safest base is a forecast that keeps your seasonal shape and takes its growth from comparable weeks. Luca's revenue forecast is built from your order history and allows for seasonality and holidays. It is checked against actuals every night. The [revenue forecast page](https://ask-luca.com/product-docs/forecasting/revenue-forecast) covers how the forecast itself works; this page covers how to use it for next quarter's numbers.

## Read next quarter in the revenue forecast {#read-next-quarter}

Here is the sequence a founder follows in November to read the quarter after peak.

1. Open **Numbers** and the **Forecast** tab.
2. Find the **Revenue Forecast** card. It shows a weekly chart of actual revenue, the same weeks last year and a shaded forecast.
3. Scroll the chart past peak into January, February and March. Note where the forecast sits against last year in each month.
4. Expand **How it's calculated** to see the forecast's assumptions, such as seasonality, growth, promotions, ad spend and stock.
5. Write down the quarter's forecast total and the growth it implies against last year.

![The Revenue Forecast card on the Forecast tab, with a weekly chart showing actual revenue, the same weeks last year and a shaded forecast that peaks at Black Friday and falls into January.](https://ask-luca.com/product-docs/images/use-cases/revenue-forecast-q1.webp)

*The chart runs past peak into January, so you can see the dip against last year before setting targets. The figures are sample data for a demo brand.*

What goes wrong at this step is reading the peak rather than the quarter. The tallest bars on the chart are the ones you are not planning for. Look at the weeks after the holiday line, and at the gap between the forecast and last year in those weeks.

> **Try asking Luca:** "What is the revenue forecast for January, February and March, and how does each month compare with last year?"

## Ask what drives the number {#ask-what-drives-it}

A forecast you cannot explain is hard to defend in a team meeting. Before you turn it into targets, find out what it assumes.

Start with the assumptions under **How it's calculated**. Then ask follow-up questions in chat, which answers across every connected platform in plain English. What-if questions work there too, and Luca reasons from your own data when it answers them.

> **Try asking Luca:** "Why is the January forecast lower than December, and how much of next quarter's growth comes from returning customers?"

Luca also holds your goals, constraints, cost rules, lead times and seasonality as business context. It follows your own order of trust between sources when it answers. This is where a founder records that spring outerwear lands late, or that cash must never drop below a set floor.

![The Context panel with the Business context section open, listing goals, rules and constraints, supplier lead times and seasonality, and an order of trust between sources.](https://ask-luca.com/product-docs/images/use-cases/business-context.webp)

*Business context gives the forecast's answers your own rules and lead times, which shape a sensible stock budget. The figures are sample data for a demo brand.*

If the answer involves ad spend or your accounts, [cross-source answers in chat](https://ask-luca.com/product-docs/chat/cross-source-answers) shows how one question covers Shopify, ads and Xero together. For questions about which customers come back after peak, see [customer retention analysis](https://ask-luca.com/product-docs/use-cases/customer-retention-analysis).

## Set next quarter's targets and the stock budget {#set-targets}

With the forecast read and explained, you can set two numbers: what the team aims for and what the business commits to buying.

### Revenue targets for the team {#revenue-targets}

Use the forecast as the base, not the target. A common approach is to set the team target a modest stretch above the forecast, and plan costs on the forecast itself. Split the quarter by month so January is judged against January.

Luca tracks goals and pacing to plan through forecasting and the control centre. As the quarter runs, you see whether each month is pacing ahead of or behind the plan. Keep the KPI definitions fixed so the team's target and the reported figure mean the same thing; [KPI reporting](https://ask-luca.com/product-docs/use-cases/kpi-reporting) covers that.

### The stock budget {#stock-budget}

The stock budget should follow the forecast, not the stretch target. If the team beats the base, you reorder; if it misses, you are not left holding a quarter's worth of extra stock.

Open the **Stock** tab to see stock cover, run-out dates and order-by dates per product. Luca also forecasts demand and stock by SKU, with seasonality built into stock planning. The [inventory demand forecasting guide](https://ask-luca.com/product-docs/use-cases/inventory-demand-forecasting) covers product-level demand. You choose the quantities and place the orders yourself.

If you connect Xero or QuickBooks, the **Cash** tab adds a cash-flow forecast up to 12 months ahead. That shows whether the stock budget leaves enough cash through the slow weeks.

### This season's context for apparel brands {#season-context-2026}

Weather is one thing your order history cannot forecast. In its retail sales bulletin covering September 2026, the Office for National Statistics reported that [online clothing retailers said autumn and winter ranges were hit by warmer weather](https://www.ons.gov.uk/businessindustryandtrade/retailindustry/bulletins/retailsales/latest). The same release estimated that UK retail sales volumes fell 0.4% in September 2026. These are estimates and the ONS revises them.

For an apparel brand, that is a reason to compare the forecast with the last few weeks of actuals before you lock the stock budget. If recent weeks ran below forecast for a reason outside your data, set the stock budget below the forecast too.

## How to check a sales forecast against actuals {#check-against-actuals}

A forecast earns trust by how close it came to what then happened. Two measures cover most of what a founder needs.

### Error in pounds {#error-in-pounds}

The mean absolute error, or MAE, is the average gap between forecast and actual, ignoring sign. The [forecasting textbook](https://otexts.com/fpp3/toolbox.html) notes that MAE is popular because it is easy to understand and compute. Its weak spot is scale: the same error in pounds means more in a slow January than in a busy December, so compare MAE month for month.

### Error as a percentage {#error-as-a-percentage}

A percentage error divides the gap by actual sales. It lets you compare a small month with a big one. Its weak spot is the slow month itself: a small denominator makes January errors look large, even when the pounds are modest.

### Judge it on months it had not seen {#genuine-forecasts}

Accuracy only counts on new data, meaning forecasts made before the sales happened. A model that fits past data closely can still forecast poorly. So judge a forecast by the forecasts it made at the start of each month, not by how well it redraws history.

### Where Luca shows it {#where-luca-shows-it}

Luca measures forecast accuracy against actuals every night. The forecasts run on a daily schedule and are compared with what actually sold. Before you set targets, compare the recent months' forecasts with what happened. If the forecast has run high, plan the quarter a little lower.

![The Forecast accuracy table comparing each month's forecast, made at the start of the month, with the actual revenue and the percentage difference from July to October.](https://ask-luca.com/product-docs/images/use-cases/forecast-accuracy.webp)

*Recent months' forecast against actual shows how far to trust next quarter's number. The figures are sample data for a demo brand.*

## How to create a sales forecast for an ecommerce brand by hand {#how-to-create-a-sales-forecast}

You can build a defensible quarterly forecast in a spreadsheet. The steps below are the manual version of what this page has described.

### The manual method {#manual-method}

1. Export at least two years of orders from Shopify and total revenue by month, net of refunds.
2. Mark one-off months, such as a stock-out or a viral product, and note how far they sit from normal.
3. Take last year's revenue for each month of next quarter. That is your seasonal base.
4. Work out year-on-year growth from comparable, non-peak months, such as September and October.
5. Apply that growth to each month of the seasonal base, not to the quarter as a whole.
6. Adjust for known changes: planned promotions, price changes and new products.
7. Set the team target as a stretch above the base, and the stock budget on the base.
8. Each month, compare the forecast you made at the start of the month with actual sales, and adjust the next month.

The weak points are steps 4 and 8. Growth chosen from the wrong weeks inflates the whole quarter, and few teams keep up the monthly check.

### The same job in Luca {#same-job-in-luca}

In Luca, steps 1 to 5 are the **Revenue Forecast** card. It is built from your full Shopify history, allows for seasonality and holidays, and plots the same weeks last year beside the forecast. Step 6 is the assumptions list under **How it's calculated**, plus follow-up questions in chat. Step 8 happens every night without you. You still own step 7: the targets and the stock budget are your decisions.

### A worked example: next quarter's targets after peak {#worked-example}

This is an illustrative calculation with sample figures, not a Luca output. A Shopify apparel brand starts from these inputs:

| Input | Sample figure |
|---|---|
| Last year's January, February and March | £60,000, £70,000 and £90,000 (£220,000 in all) |
| Average peak week this year | £30,000 |
| Growth on last year in November | 25% |
| Growth on last year in September and October | 10% |
| Stock cost per £1 of revenue | 40p |

Each method then gives a different quarter and a different stock budget:

| Method | Arithmetic | Next quarter | Stock budget at 40% |
|---|---|---|---|
| Recent run rate | 13 weeks × £30,000 | £390,000 | £156,000 |
| Last year plus peak growth | £220,000 × 1.25 | £275,000 | £110,000 |
| Last year plus steady-month growth | £220,000 × 1.10 | £242,000 | £96,800 |

| Comparison | Arithmetic | Result |
|---|---|---|
| Extra stock tied up by the run-rate method | £156,000 minus £96,800 | £59,200 |
| Team target with a 5% stretch on the steady-month base | £242,000 × 1.05 | £254,100 |

The run rate carries peak into January and buys stock the quiet weeks will not sell. The steady-month row is the base to plan stock on. A team target about 5% above it gives the team a stretch without buying stock for it.

## Get forecasting switched on {#switch-on}

Forecasting is switched on per client by Luca's team, and it is not gated by plan. Ask your Luca contact to switch it on for your brand. Once it is on, the **Forecast** tab appears under **Numbers** and the forecast refreshes daily.

The free-cash-flow forecast needs Xero or QuickBooks connected. If you sell on Amazon, Amazon Seller Central brings in up to two years of history alongside your Shopify orders.

## Who uses ecommerce sales forecasting, and how {#who-uses-it}

### Founders and CEOs {#founders}

A founder uses the forecast to set next quarter's targets and to approve the stock budget behind them. The question is whether the plan holds up after peak. For a daily view of how sales, ads and cash are moving, see [seeing the whole business in one place](https://ask-luca.com/product-docs/use-cases/see-the-whole-business).

### CFOs and finance leads {#finance-leads}

A finance lead checks the forecast's accuracy before it goes into the budget, then tests whether the stock budget leaves enough cash. With Xero or QuickBooks connected, the cash-flow forecast sits next to the revenue forecast.

### Growth and marketing leads {#growth-leads}

A growth lead uses the forecast to judge how hard to push spend in a slow month. Before moving budget into January, [ROAS analysis](https://ask-luca.com/product-docs/use-cases/roas-analysis) shows which channels have been earning back their spend.

## How Luca AI compares for e-commerce sales forecasting {#compare}

Luca AI builds its forecast from your own Shopify orders, allows for seasonality and holidays, and checks the forecast against actuals every night. Several tools cover one of these; the grid shows which.

E-commerce sales forecasting features, from each platform's public documentation:

| Platform | Built from your own order history | Allows for seasonality and holidays | Forecast checked against actuals nightly |
|---|---|---|---|
| **Luca AI** | Yes | Yes | Yes |
| **Lifetimely by AMP** | Yes (from historical Income Statement sales) | [Does not apply seasonality](https://help.useamp.com/article/1389-lifetimely-forecasting) | Not in docs |
| **Daasity** | Partly ([you input projected sales](https://help.daasity.com/core-concepts/custom-data-and-enrichment/revenue-plan-forecast/e-commerce-revenue-plan)) | Partly ([you plan daily around events](https://help.daasity.com/core-concepts/custom-data-and-enrichment/revenue-plan-forecast)) | Partly ([daily plan against budget projections](https://help.daasity.com/core-concepts/data-models/data-explores/digital-analytics-unified/forecast-vs-actual-ecommerce-explore)) |
| **Shopify Sidekick** | Partly ([writes code to forecast on request](https://help.shopify.com/en/manual/ai-powered-tools/sidekick)) | Not in docs | Not in docs |
| **Triple Whale** | Partly ([suggests forecasting from order data](https://triplewhale.readme.io/docs/non-zero-orders)) | Not in docs | Not in docs |

*Yes: documented and built in. Partly: documented with a limit, shown in brackets. Not in docs: not found in the vendor's public documentation.*

### Why Luca AI is the stronger choice for e-commerce sales forecasting {#compare-stronger}

- **Seasonality is part of the forecast.** Luca allows for seasonality and holidays such as Black Friday, while Lifetimely's forecasting [does not match January to January or apply seasonality](https://help.useamp.com/article/1389-lifetimely-forecasting). See [Luca AI vs Lifetimely](https://ask-luca.com/blogs/lifetimely-alternatives).
- **The forecast comes from your orders, not your inputs.** Luca builds the forecast from your Shopify history, while Daasity's revenue plan is where you [input your projected sales](https://help.daasity.com/core-concepts/custom-data-and-enrichment/revenue-plan-forecast/e-commerce-revenue-plan). See [Luca AI vs Daasity](https://ask-luca.com/blogs/daasity-alternatives).
- **The forecast is ready before you ask.** Luca forecasts daily revenue up to a year ahead and checks it every night, while Triple Whale's docs [suggest using order data to forecast](https://triplewhale.readme.io/docs/non-zero-orders). See [Luca AI vs Triple Whale](https://ask-luca.com/blogs/triple-whale-alternatives).

*Competitor details checked against each vendor's public documentation on 7 November 2026. Features and plans change, so confirm with the vendor before you decide.*

## Limits and things to know {#limits}

- **Ecommerce revenue, not B2B pipeline.** The forecast is built from store orders. It does not roll up CRM deals by stage or close date.
- **Forecasting is switched on per client.** Luca's team switches it on for your brand.
- **The free-cash-flow forecast needs Xero or QuickBooks.** Without an accounting connection, you get revenue and product demand forecasts.
- **Read-only by design.** Luca never changes your store, ads or books, so targets, orders and spend changes stay your decisions.
- **Data refreshes overnight.** Today's orders appear in the forecast after the next overnight refresh.

## Common questions {#common-questions}

#### How do I create a sales forecast for an ecommerce brand? {#how-do-i-create-a-sales-forecast}

Start from your own order history. Take last year's revenue for each month of next quarter, apply growth from comparable non-peak months, and adjust for known changes. In Luca, the revenue forecast on the **Forecast** tab does this from your Shopify history, with seasonality and holidays allowed for.

#### Is this the same as B2B sales forecasting? {#is-this-b2b-sales-forecasting}

No. B2B sales forecasting rolls up deals in a CRM pipeline by stage and close date. Ecommerce sales forecasting starts from orders your store has already taken and the seasonal pattern they show. Luca's revenue forecast is built from your store's orders, not from CRM deals.

#### How much order history do I need for a quarterly forecast? {#how-much-history}

You need at least one full year, so next quarter has last year's version to compare with. Two or more years is better, because one odd year can mislead. Seasonal models such as Prophet [work best with several seasons of history](https://cran.r-project.org/web/packages/prophet/index.html). Luca imports your full Shopify history.

#### Why does the January forecast look so low next to December? {#why-is-january-low}

Because January is usually low for seasonal brands. A forecast built from your history keeps the dip that follows peak. Compare January's forecast with last January on the **Revenue Forecast** card, not with December. If January beats last January, you are growing.

#### Should my team's target equal the forecast? {#target-equal-forecast}

Usually not. Use the forecast as the base for stock and costs, then set the team target a modest stretch above it. That way, beating the target is a win, and missing it does not leave you holding extra stock or short of cash.

#### How do I know how far to trust the forecast? {#how-far-to-trust}

Look at how close recent forecasts came to actual sales. Luca checks every forecast against actuals each night and corrects it. If the recent months show the forecast running high, plan the quarter a little lower than the forecast says.

#### Can I compare the forecast with last year? {#compare-with-last-year}

Yes. The **Revenue Forecast** card plots actual revenue, the same weeks last year and the shaded forecast on one weekly chart. The gap between the forecast and last year is the growth the forecast assumes, which is the number to question before you set targets.

#### Does the forecast include my planned promotions? {#planned-promotions}

Expand **How it's calculated** on the **Forecast** tab to see the forecast's assumptions, such as seasonality, growth, promotions, ad spend and stock. If a planned promotion or price change is missing, adjust your target for it and ask about its effect in chat.

#### Can I ask what happens if I cut ad spend in January? {#cut-ad-spend}

You can ask what-if questions in chat, and Luca reasons from your brand's own data when it answers. Use the answer as a starting point for the target conversation, then compare it with recent months' actuals before you commit.

#### How does the sales forecast help with the stock budget? {#stock-budget-help}

Base the stock budget on the forecast, not on the stretch target. The **Stock** tab shows stock cover, run-out dates and order-by dates, and Luca forecasts demand by SKU with seasonality built in. You choose the quantities and place the orders yourself.

#### Can I forecast cash as well as revenue? {#forecast-cash}

Yes, if you connect Xero or QuickBooks. The free-cash-flow forecast needs one of them. The **Cash** tab then shows a cash-flow forecast up to 12 months ahead, so you can check whether the stock budget leaves enough cash through a slow January.

#### Does Luca change anything in my Shopify store? {#does-luca-change-my-store}

No. Luca only reads your Shopify data and never changes your store. It writes to none of the systems you connect, so targets, stock orders and spend changes stay your decisions.

#### How often does the forecast update? {#how-often-update}

Forecasts run on a daily schedule. Every connected source refreshes once a day, overnight, and the forecast is compared with actuals every night. Orders taken today show up after the next overnight refresh.

#### How do I get forecasting switched on? {#get-it-switched-on}

Forecasting is switched on per client by Luca's team and is not gated by plan. Ask your Luca contact to switch it on for your brand. The **Forecast** tab then appears under **Numbers**.

## Further reading {#further-reading}

- [Forecasting: Principles and Practice, The forecaster's toolbox](https://otexts.com/fpp3/toolbox.html): the seasonal naive and naive methods, MAE and percentage errors, and why accuracy is judged on new data.
- [Forecasting: Principles and Practice, Exponential smoothing](https://otexts.com/fpp3/exponential-smoothing.html): how Holt-Winters models estimate level, trend and seasonality from history.
- [Prophet on CRAN](https://cran.r-project.org/web/packages/prophet/index.html): a seasonal forecasting model with holiday effects, and the history it needs.
- [ONS retail sales bulletin](https://www.ons.gov.uk/businessindustryandtrade/retailindustry/bulletins/retailsales/latest): the latest UK retail and online sales estimates, including clothing.

## Related {#related}

- [Shopify revenue forecast](https://ask-luca.com/product-docs/forecasting/revenue-forecast): how the forecast is built, read and checked every night.
- [Ecommerce demand forecasting](https://ask-luca.com/product-docs/use-cases/inventory-demand-forecasting): forecast each product's demand before you reorder.
- [AI agents for ecommerce](https://ask-luca.com/product-docs/use-cases/ai-agents-for-ecommerce): what an AI analyst does on your store data and what you decide.
- [Conversational analytics across sources](https://ask-luca.com/product-docs/chat/cross-source-answers): ask one question across Shopify, ads and Xero.
- [Customer retention analysis](https://ask-luca.com/product-docs/use-cases/customer-retention-analysis): which cohorts and products keep customers.
- [KPI reporting](https://ask-luca.com/product-docs/use-cases/kpi-reporting): one set of KPI definitions in dashboards, chat and exports.
- [ROAS analysis](https://ask-luca.com/product-docs/use-cases/roas-analysis): ROAS by channel and campaign against break-even ROAS.
- [See the whole business](https://ask-luca.com/product-docs/use-cases/see-the-whole-business): sales, ads, profit and cash in one place each morning.
- [Set up Luca](https://ask-luca.com/product-docs/getting-started/set-up-luca): connect sources and invite your team.
- [Read-only access](https://ask-luca.com/product-docs/platform/read-only-access): how Luca's read-only access works.
- [Luca AI product documentation](https://ask-luca.com/product-docs): every guide in one place.

*Updated 7 November 2026: first published, with forecasting methods for targets after peak, an accuracy check, a manual method, a worked example and 14 common questions.*