This article contains affiliate links. AutomateWithMNK may earn a commission if you sign up through certain links, at no additional cost to you. This does not influence our editorial assessment.
A recurring CRM export needs a defined dataset, authorised connections, a destination and a checked first run. This documentary tutorial follows Coupler.io’s HubSpot Deals source into Google Sheets. It creates a current opportunity report, rather than moving the sales process into a spreadsheet.
The example is a fictional equipment supplier, Cedar Works. Its manager wants a repeatable view for a weekly pipeline meeting. This is different from a one-time exit package before cancelling a CRM or an import into a replacement system.
Disclosure: this article includes a Coupler.io affiliate link. We may earn a commission if you purchase through that link.
Documentary analysis dated 14 October 2026, without a product account test. Numerical examples are fictional and illustrate the method.
Write the extraction contract
Cedar’s report includes open opportunities from one agreed pipeline. One row represents an opportunity, not a person or an email interaction. The representative remains responsible for updating the opportunity in the CRM.
Prepare a workbook with separate raw-data and analysis tabs. Identify the Google account allowed to edit it and the person who will maintain the connection. Avoid beginning in a workbook containing valuable formulas inside the intended import area.
Record the expected CRM portal as well. An operator with access to several portals can select the wrong one and still receive plausible data. The report contract should identify the business scope, its owner and the records excluded from the first version. A narrow, explainable extraction is easier to accept than an unrestricted download.
Choose fields that describe one opportunity
| Reporting field | Example | Meaning |
|---|---|---|
| Opportunity ID | CW-401 | Stable identifier from the source |
| Pipeline stage | Proposal | Stage in the selected pipeline |
| Amount | 6 200 GBP | Numeric value with explicit currency |
| Owner | Jamie | Person responsible for follow-up |
| Expected close | 2026-11-06 | Forecast date, not cash receipt |
| Extraction time | 2026-10-14 06:00 UTC | Age of the reporting copy |
These are reporting labels, not a promise of specific API field names. Map them to the technical properties available in your authorised source.
Configure the documented HubSpot source
Select Add data flow, find HubSpot and use Sign in with HubSpot. Choose the intended portal, authorise the connection and select Deals. This sequence follows the official HubSpot source documentation.
Select only the columns the report needs. For custom properties, check the internal field name against the preview rather than assuming the familiar display label is the technical name. Preserve the source identifier even if you later give other columns friendlier reporting labels.
Be cautious with date filters. Cedar’s current pipeline includes an old opportunity that is still open. A recent-created filter would omit it. Record any filter’s purpose and inspect boundary cases before making it part of the recurring report. Filtering for convenience can silently change the question the report answers.
Connect Google Sheets and choose the writing area
Choose Google Sheets as the destination, use Connect for the authorised Google account and select the existing workbook and target tab. The destination documentation explains this part of the setup.
Cedar writes into opportunities_raw and keeps its calculations in pipeline_view. A representative’s annotation belongs in a separate controlled area, because it should not disappear when the raw dataset is replaced. This is a reporting design choice, not a claim that the service merges human comments automatically.
Check the starting cell and any explicit range boundary. A small bounded range can make a sample look tidy while excluding rows from a larger extraction. Use a dedicated area that accommodates the expected data, then verify how many identifiers and fields arrived. Review destination access before adding sensitive columns.
Decide whether you want current state or observations
Cedar wants the latest accepted state of its pipeline. If CW-401 changes stage, the current report should describe that change without presenting two current opportunities. Replace is the relevant writing mode to investigate for the dedicated raw area.
Append serves a different purpose: adding observations. Repeatedly appending the entire current opportunity list creates repeated copies unless your design explicitly distinguishes them. It does not supply automatic uniqueness merely because each run completes successfully.
If management later requests historical pipeline movement, define a separate observation date and retention process. A current-state extract does not reconstruct every previous stage. Keep that limitation visible so the weekly report is not later treated as an archive it was never designed to provide.
Run manually and reconcile more than the total
Use Run for the first extraction. Compare the output with the same pipeline and inclusion rules in the CRM. A native report may apply a different time range or ownership filter, so explain the comparison before declaring either result incorrect.
Trace several known opportunities individually. Check the identifier, stage, owner, amount and currency. A matching row count can hide one missing record and one duplicate record. Missing values should remain identifiable rather than becoming invented zero amounts.
Save the accepted sample and run the flow again without changing the source. A current-state report should remain understandable and should not double solely because it ran twice. Investigate unexpected changes before enabling unattended operation. The second run is a practical check of the report design, not a performance claim about the product.
Use a filled acceptance checklist
| Case | Expected output | What to inspect |
|---|---|---|
| Old but open CW-401 | Still included | Creation-date filters |
| CW-402 has no amount | Missing value visible | Numerical transformations |
| CW-403 closes | Leaves open-opportunity view | Stage inclusion rules |
| Repeat unchanged run | Same current opportunity set | Identifiers and writing mode |
| A new source property appears | Stable report columns | Field selection |
| Authorisation withdrawn | Old dataset clearly dated | Run history and maintenance owner |
Enable the useful refresh cadence
After accepting the manual results, configure automatic refresh in the flow settings. The refresh documentation ties available intervals to the selected plan. Cedar sets a daily reporting requirement for its weekly meeting; this requirement is not an assurance that every run will succeed.
Display the time of the last useful extraction in the reading view. An unchanged opportunity does not prove the flow has stopped, but the reader should be able to distinguish recent information from an old copy. Treat the timestamp as part of the report’s meaning.
Name a maintainer and backup person. They should know the flow, authorising accounts and intended destination. Agree when they notify the manager after a failure. Do not let an apparently plausible table keep implying freshness while a disconnected account prevents new data from arriving.
Keep the spreadsheet’s authority limited
Representatives continue to update opportunities in the CRM. Cedar’s spreadsheet supports reading and analysis. Turning it into another place to modify sales records would require a separate write-back design and clear ownership rules.
An export can reflect duplicates already present in the source. Our HubSpot and Make duplicate-prevention guide examines the different job of safely creating contacts. A reporting flow should not be mistaken for a remedy for every operational data-quality problem.
For each proposed new column, ask which decision it improves. Keep unnecessary personal details out of the report. Expand the field dictionary only when someone can explain the new information and accept its access requirements. This makes later troubleshooting considerably more concrete.
FAQ : practical questions
Can I reuse these steps for every CRM?
No. The source procedure is documented for HubSpot. Confirm the connector, entities and authorisation flow for another CRM before assuming that the same configuration applies.
Why might old opportunities disappear?
Review date filters and the authorising account’s permissions. A current pipeline view may need opportunities created long before the reporting period.
Does Append remove duplicates?
No. It adds rows. Define the extraction boundary and an observation key if you need history, then compare consecutive runs to understand what is being added.
Is this a complete CRM backup?
No. This report selects fields for an analytical view. A cancellation or restoration package needs a separate inventory of objects, relationships, documents and retention requirements.
Choose your next step
Create the field dictionary and a dedicated destination, then reconcile the first two manual runs. Enable automatic refresh only after the result is accepted. If Coupler.io fits the source and cadence, the affiliate link below opens its website for evaluating the relevant offer.
Evaluate Coupler.io for a recurring CRM export
Official sources
Checked on 14 October 2026.
