Export Google Ads to JSON Without Breaking Dates
Brand promotion: this article introduces Guangsuan's Google Ads management service. It is a vendor introduction, not an independent review.
Short answer: if you need Google Ads data in JSON for month-end reconciliation, do not look for a JSON button in the reporting interface. There is none. JSON comes from the Google Ads API REST interface, or from a warehouse or connector that exports JSON downstream. Before either file is used for attribution, lock the date range and account time zone, the conversion basis and counted actions, and the currency unit for cost. A file that parses as JSON is not yet a file that can be reconciled.
There is no JSON button in the interface
When you click the download icon above a Google Ads statistics table and choose Download, the listed formats are Excel CSV, TSV, PDF, XLSX, XML, and Google Sheets. JSON is not one of them. XML is the closest native format, but XML and JSON are different languages for different consumers.
If a downstream tool insists on JSON, converting a downloaded CSV or XML file locally is a programming step you own, not an export option. Keep that script versioned, name an owner, and test it on a small slice checked by hand. For the full format comparison and API route, see Guangsuan's guide to exporting Google Ads data to JSON.
What each route solves
Choose the route by what happens next.
- Native download for reading: On the Campaigns, Ad groups, Keywords, Search terms, or Ads page, set the date range and columns, then download. Use the Reports area when day, device, geographic, or conversion splits must sit in one table. Use CSV for scripts and databases, XLSX for spreadsheet work, and Sheets when several people need the same live file.
- REST API for JSON: The Google Ads API supports gRPC and REST interfaces. The REST interface uses JSON representations of the same resources. Queries are written in Google Ads Query Language, and a GoogleAdsService query returns rows carrying the attributes, metrics, or segments you selected. This route solves automation and typed JSON output, but it requires a developer token, an authenticated client, and credential storage.
- Warehouse or connector for recurring joins: A BigQuery transfer or reporting connector can hold the data and emit JSON downstream. This route solves history and repeatable CRM matching, but it adds a vendor, project, or scheduled transfer to maintain. For a one-off file, local CSV-to-JSON conversion is often faster with no vendor attached.
Access control matters on the API route. A developer token is required, and its access level sets daily operation limits. A new project starts at Test Account Access for test accounts before you apply for a higher level, and permissible use decides which features you may call. Keep the token and credentials in a secured file or secret store, never in a public report.
What the JSON shape looks like
What can be shown honestly before you have access is the shape, not live account data. The example below does not call the API and contains no credentials. It is a hypothetical teaching illustration with invented figures.
import json
rows = [
{"campaign.id": "1234567890", "campaign.name": "Brand-Exact", "metrics.clicks": 412, "metrics.conversions": 18, "metrics.cost_micros": 26140000},
{"campaign.id": "1234567891", "campaign.name": "Generic-Long-Tail", "metrics.clicks": 903, "metrics.conversions": 11, "metrics.cost_micros": 88750000}
]
text = json.dumps(rows, indent=2)
print(text)Two details decide whether this shape joins correctly. First, keys use dotted names such as campaign.name and metrics.clicks because they map directly to the GAQL fields selected. If campaign ID, date segment, or conversion action is missing from the query, no downstream join restores it. Second, cost arrives as metrics.cost_micros in account currency. Divide once by one million in a noted place, and do not normalize an already normalized amount again.
Lock dates, conversions, and currency before reconciliation
An export is only as good as the range and time zone behind it. Record both with the file.
- Range and grain: A monthly export can hide a weak week, while a daily export over a long range becomes hard to review. Export the grain you will actually analyse. Use the segment control to split by day, week, or month rather than joining several separate ranges later. Keep customer ID, campaign ID beside campaign name, and every split that defines a row, such as date, device, and conversion action where used. Names are edited; IDs provide the stable key.
- Time zone: Exports are dated in the account time zone, which may not match your calendar or CRM timestamps. A lead recorded late on the last day of the month can fall into different months on the two sides with no error on either side. Write the zone on the file and check one date row after opening.
- Conversion basis: Google Ads can report conversions on more than one basis. Conversion-time reporting counts a conversion in the period when it happened; click-time reporting counts it in the period when the click happened. A click on 30 September that becomes a 7 October signup sits in different months under the two rules. Neither basis is wrong, but mixing them in one comparison measures the setting rather than performance. Pick one basis and record the conversion window, counting method, and attribution model beside the file.
- Which actions count: Only conversion actions marked as primary feed the Conversions column used for bidding. Secondary actions, such as a page view or click-to-contact micro-conversion, are reported separately and should be labelled as such. Changing the primary set, counting, or model changes historical comparisons.
- Cost and files: Add the cost column and compute ratios in the sheet rather than relying only on exported averages. Keep the raw download unchanged, work in a copy, name the copy with account, range, basis, and pull time, and replace the same slice on retry rather than appending a duplicate.
Reconcile against attribution in a fixed order
An export from Google Ads shows what Google was able to record. Sales-confirmed numbers live in your own system. Keep clicks, platform conversions, qualified inquiries, and closed revenue as four separate counts. Adding them produces a number that describes nothing.
Run one narrow pass before touching bids:
- Select one primary action, one campaign group, and one closed period plus the first days of the next month to catch late entries.
- Align range, time zone, and conversion basis on both extracts.
- Separate ad-sourced records from organic and referral entries.
- Check counting settings and repeated uploads before calling a remainder unexplained. Very quiet rows can drop from a filtered view, so confirm filters and grain before saying a term never matched.
- Leave any remaining gap marked as open rather than assigning unknown traffic to Ads.
As a hypothetical teaching illustration with invented figures, not a client result: a file records 52 ad-tracked contacts in a month. Sales accepts 17 as qualified under a written definition, with 12 dated inside that month and 5 in the next month, and 2 reach order. Each total describes a different stage — platform recording, commercial validation, and business outcome — so none contradicts another.
Where Guangsuan fits
Guangsuan provides Google Ads management covering account structure, search terms, landing pages, conversion checks, and reporting. That scope matters here because some reconciliation causes sit in the account, not the file: a conversion action that records the wrong step, a landing page that does not match the query, or tracking that was deployed and never checked. Details, fees, and handoff are confirmed in a contract, with no fixed inquiry-volume promise. See Guangsuan Google Ads management.
For a first review, bring the conversion-action list with primary status marked, plus one real export with its date basis and time zone written on it.
FAQ
Can I download JSON directly from Google Ads?
No. Use the REST API path or a warehouse path, or convert a one-off CSV locally with a versioned script.
Does a file that parses as JSON reconcile cleanly?
No. Parsing says nothing about grain, micros, time zone, conversion basis, or which actions were summed.
Click time or conversion time for month-end?
Use the basis that answers the question and keep it on both sides, with window and account time zone filed with the extract.
Ready to turn this into pipeline?
Book a 30-minute paid strategy audit with a senior strategist. We'll map your current spend, find the leaks, and model the 90-day lift.
Get My Free Strategy Audit