Spendwise field guide

From first sign-in to a forecast you can actually inspect.

Follow this end-to-end walkthrough to create your workspace, add campaign history, connect ad platforms, understand each dashboard view, and validate the model against what happened next.

01

Set up

Create a workspace and choose a project.

02

Connect

Import a CSV or sync an ad platform.

03

Understand

Read forecasts and test past accuracy.

01 / Set up your account

Create the place where your campaign history will live.

Spendwise keeps data, forecasts, reports, and permissions scoped to a project inside your workspace.

1

Sign in with Google

Open Spendwise and choose Continue with Google. No Spendwise password is created.
2

Choose a workspace

Create a private personal workspace, create an organization for a team, or accept an invitation that is already waiting for your Google account.
3

Name your first project

Complete onboarding with your name, role, workspace type, and project details. You can add more projects later from Settings.

02 / Bring in campaign data

Choose the source that fits your workflow.

After onboarding, open Settings → Project integrations. Select a source, save it to the selected project, and use Sync now whenever you want to pull fresh history.

Google Ads

Add the OAuth client details and account information for the Google Ads project you want to analyze.

  • Customer ID is required
  • OAuth client ID, secret, and refresh token are used for the connection
  • Optional manager customer ID and developer token fields are available
  • Save the connection, then select Sync now

Meta Ads

Add the Meta app credentials and long-lived access token for the ad account you want to sync.

  • Ad account ID is required
  • App ID, app secret, and long-lived access token are supported
  • Save the connection to the current project
  • Select Sync now to retrieve campaign history

CSV import

Use Import CSV from Overview when you already have campaign history in a spreadsheet.

  • CSV files can be up to 10 MB
  • Required: campaign name, date, and spend or conversions
  • Common column names are matched automatically
  • Revenue or conversion value enables revenue-based ROAS

CSV quick checklist

  • Campaign name
  • Date
  • Spend or conversions
  • Revenue or conversion value for ROAS
  • Campaign ID, source, status, impressions, and clicks are optional

If a connection saves but does not sync, check the account ID, token permissions, and the last sync message in Project integrations. A saved connection does not populate the dashboard until a sync succeeds.

03 / Read the dashboard

Use Overview to see what changed and where to look next.

A

Metric cards

Review blended CPA, conversions, spend, and ROAS for the selected date range. Use the range control to change the period.
B

Trend chart

Compare recent campaign performance over time. Empty states explain when the project needs data before a chart or forecast can be calculated.
C

Campaign table

Scan campaign-level performance, filter by source, and open Campaigns for the full library and more detailed filtering.
D

Next actions

Use Generate forecast, Run backtest, or Import CSV from the Overview actions. If data is missing, the dashboard points you back to Settings.

04 / Generate forecasts

Look five days ahead, then read the range around the estimate.

The forecast workflow

Use Forecasts after a successful sync or import.

  1. 01Open Forecasts from the project navigation.
  2. 02Choose Generate forecast. Spendwise uses recent history and campaign rhythm to estimate the next five days.
  3. 03Review forecast CPA, range, direction, spend, conversions, and ROAS by campaign.
  4. 04Use the signal as a prompt to investigate. It is not an instruction or a guarantee.

How to interpret the labels

Improving: forecast CPA is meaningfully lower than recent CPA.

Stable: the change is within roughly 2.5% of recent CPA.

Declining: forecast CPA is meaningfully higher than recent CPA.

Improve, stable, and decline percentages are derived from the forecast range. They are not calibrated historical odds.

05 / Check accuracy

Measure how earlier forecasts compared with later outcomes.

Open Accuracy and choose Run backtest. Spendwise hides future periods, forecasts them, and compares those predictions with the actual history that followed.

MAPE

A measured error signal for the evaluated forecast windows.

WAPE

A measured error signal for the evaluated forecast windows.

RMSE

A measured error signal for the evaluated forecast windows.

Bias

A measured error signal for the evaluated forecast windows.

1

Check history

Backtesting currently needs at least 120 history days per campaign and eligible holdout windows.
2

Check coverage

Look at evaluated samples, rows, horizon, and history alongside the error metrics.
3

Use judgment

A measured past error helps you calibrate trust; it cannot guarantee future performance after a campaign change.

06 / Work with your team

Keep people, projects, and permissions organized.

1

Create an organization

Choose an organization workspace during onboarding when multiple people need access to shared projects.
2

Invite teammates

Open Settings, enter a Google account email, and choose the role. The invite appears in Notifications for the recipient.
3

Switch projects

Use the project switcher in the dashboard header. Campaign history, forecasts, and reports stay scoped to the selected project.

Use Notifications to accept or decline pending invitations. The available actions depend on your organization role; project and member management may be limited to owners or administrators.

07 / Settings and AI reports

Configure the project without losing the thread.

A

Project and organization settings

Settings is where you create projects, switch the active project, manage members, and review saved integrations.
B

AI analysis

Settings → AI analysis lets you choose a provider and model, add an API key, and save the configuration for the selected project.
C

Generate the report

Forecasts → Generate AI report turns available campaign signals into written drivers and suggested checks. AI explains the forecast; it does not calculate it.
D

Know when to skip AI

Forecasts and backtests work without an AI key. Add AI only when a written explanation helps your workflow and your provider configuration is ready.

Quick troubleshooting

When a view looks empty, check these first.

No campaigns appear

Confirm that the import or Sync now completed for the selected project, then check the source filter in Campaigns.

Forecast is unavailable

A campaign needs at least 21 active history days with enough spend and conversion data for a forecast.

Backtest has no result

Add more history. The current backtest needs at least 120 history days per campaign and eligible forecast windows.

AI report cannot be generated

Save a provider, model, and API key under Settings → AI analysis, then return to Forecasts.

A teammate cannot see a project

Check the organization invitation, assigned role, active project, and project-management permissions.

Ready to bring in your first history?

Open your workspace and follow the setup flow. You can start with a CSV and add integrations later.

Open Spendwise