Call google shopping and amazon on the same product phrase, then read weekly shopping indexes next to monthly Amazon volume because the two calendars do not share a close.
Live data as of 2026-09-01
A Google Shopping vs Amazon search comparison on Trends MCP is two get_growth sources on the same product phrase. Source google shopping returns a weekly 0-100 shopping-intent index. Source amazon returns a monthly 0-100 index plus absolute search volume. From the September 1, 2026 pull, air fryer sat at Shopping index 8.0 on 2026-08-22, up 60.0% from 5.0 a year earlier, while Amazon sat at index 51.5 on 2026-07-31, down 30.31% from 73.9 with volume 5,577,980. Amazon 7D and 14D returned collapsed_range. Google Shopping 7D filled, and the 1W alias matched 7D on the same dates. The free plan includes 100 requests per month.
On the September 1, 2026 pull, every amazon 7D window returned error collapsed_range with the message "Recent and baseline resolved to the same data point; not enough history for this preset." Air fryer, Crocs, Owala, Nike, and running shoes all hit that error. Air fryer 14D did too. The amazon series held 49 points and closed on 2026-07-31, so a seven-day lookback cannot pick a second month.
Google Shopping 7D filled on those same keywords. Air fryer moved from 6.0 on 2026-08-15 to 8.0 on 2026-08-22 (+33.33%). Crocs moved from 49.0 on 2026-08-22 to 48.0 on 2026-08-29 (-2.04%). Owala moved from 70.0 to 73.0 (+4.29%). Nike moved from 26.0 to 28.0 (+7.69%). running shoes stayed flat at 7.0. Passing 1W next to 7D for air fryer returned the same recent_date, baseline_date, values, and +33.33%. The MCP schema lists 1W as an alias. The public llms.txt period table does not.
Short Amazon windows are a cadence limit, not a missing product. Use 30D or longer on amazon. Keep 7D for google shopping when a weekly move is the question.
Two source strings, one keyword. REST uses mode get_growth. MCP uses the get_growth tool. The shopping-only page is the Google Shopping trends API. The Amazon keyword page is Amazon search trends.
{"mode": "get_growth", "source": "google shopping", "keyword": "air fryer", "percent_growth": ["7D", "14D", "30D", "3M", "12M"]}
{"mode": "get_growth", "source": "amazon", "keyword": "air fryer", "percent_growth": ["7D", "14D", "30D", "3M", "12M"]}
{"mode": "get_growth", "source": "amazon, google shopping", "keyword": "air fryer", "percent_growth": ["7D", "12M"]}
{"mode": "get_growth", "source": "google shopping", "keyword": "air fryer", "percent_growth": ["1W", "7D"]}
Single-source shopping responses for air fryer, Crocs, Owala, Nike, running shoes, and Stanley cup listed recent_value and baseline_value with no volume_available field. Single-source amazon responses set volume_available true and printed recent_volume. metadata.total_data_points was 260 for air fryer on google shopping and 261 for Crocs, Owala, Nike, running shoes, and Stanley cup. amazon stayed at 49 for every keyword in this pull.
The comma-separated call returned a different envelope: sources_requested ["amazon", "google shopping"], sources_processed 2, sources_successful 2, plus a trend_analysis block. That is not two stacked single-source payloads. Read source_results for the per-source windows.
The September 1, 2026 12M window is the durable split. Air fryer shopping interest rose while Amazon search volume fell. running shoes did the same. Nike fell on both, but Shopping fell 12.5% while Amazon fell 23.73% with volume 14,890,482 versus 19,514,723. Owala rose on both. Stanley cup fell on both, harder on Shopping at the index (-62.5% from 8.0 to 3.0) than on Amazon volume (-54.76% from 5,004,025 to 2,263,655).
| Keyword | Shopping 12M | Shopping recent | Amazon 12M | Amazon recent | Amazon volume |
|---|---|---|---|---|---|
| air fryer | +60.0% (5.0 to 8.0) | 2026-08-22 | -30.31% (73.9 to 51.5) | 2026-07-31 | 5,577,980 vs 7,994,479 |
| running shoes | +40.0% (5.0 to 7.0) | 2026-08-29 | -29.64% (83.0 to 58.4) | 2026-07-31 | 2,930,625 vs 4,166,841 |
| Crocs | +2.13% (47.0 to 48.0) | 2026-08-29 | -20.3% (66.0 to 52.6) | 2026-07-31 | 9,187,782 vs 11,527,965 |
| Owala | +28.07% (57.0 to 73.0) | 2026-08-29 | +4.65% (75.3 to 78.8) | 2026-07-31 | 11,067,682 vs 10,580,469 |
| Nike | -12.5% (32.0 to 28.0) | 2026-08-29 | -23.73% (59.0 to 45.0) | 2026-07-31 | 14,890,482 vs 19,514,723 |
| Stanley cup | -62.5% (8.0 to 3.0) | 2026-08-29 | -54.7% (41.5 to 18.8) | 2026-07-31 | 2,263,655 vs 5,004,025 |
Shorter windows disagree with that year-long story. Air fryer shopping 3M was -84.91% (53.0 on 2026-05-23 to 8.0), while amazon 3M was +5.75% (48.7 on 2026-04-30 to 51.5) with volume 5,577,980 versus 5,272,023. running shoes shopping 3M was -86.79% (53.0 to 7.0). Those spring-to-late-summer drops sit next to a 12M print that still looks up because the year-ago shopping baseline was 5.0. Quote the window dates before calling a category recovered.
Crocs amazon 30D was -27.95% (73.0 on 2026-06-30 to 52.6) with volume 9,187,782 versus 12,765,263. Crocs shopping 30D was +6.67% (45.0 on 2026-08-01 to 48.0). Same brand, opposite 30-day direction, two different close dates.
The comma-separated air fryer call returned volume_weighted_growth -30.23 with has_volume_data true, sources_with_volume 1, weight_coverage "1/2 sources", and total_volume 5,577,980. The volume_details row named amazon at 100.0% of the weight and used amazon volume_growth -30.23, not the amazon index growth -30.31. Google Shopping contributed 0 volume, so it could not pull the blended number toward the shopping 12M of +60.0%.
trend_analysis printed uptrend_count 1, downtrend_count 1, uptrend_ratio "1/2". That count mixes a shopping 7D increase (+33.33%) with an amazon 12M decrease, and amazon 7D never calculated. An uptrend ratio on mixed windows is a tally of signed prints, not a shared month. For the envelope fields themselves, see the multi-source growth API.
Live Amazon Best Sellers ranks are a third object. They use get_top_trends type Amazon Best Sellers Top Rated or Amazon Best Sellers by Category, not the amazon keyword source. Join a rank to these growth series in code. Do not expect one JSON list to hold both.
amazon recent_date was 2026-07-31 on every keyword in this pull. google shopping recent_date was 2026-08-22 for air fryer (datetime form 2026-08-22 00:00:00+00:00, 260 points) and 2026-08-29 for Crocs, Owala, Nike, running shoes, and Stanley cup (261 points). A ratio that divides today's shopping index by last month's Amazon volume invents a week that neither series published.
30D baselines show the same clock split. Shopping 30D for Crocs, Owala, Nike, and running shoes used baseline_date 2026-08-01 against recent_date 2026-08-29. Amazon 30D used 2026-06-30 against 2026-07-31. Those are not 30 days apart on one calendar. They are adjacent months on amazon and a four-week shopping span that never touches July 31.
Air fryer shopping 30D used 2026-07-25 against 2026-08-22 (+14.29%, 7.0 to 8.0). That is the one shopping keyword in this set whose recent week lagged the August 29 close. Treat metadata.total_data_points and recent_date as part of the answer, not footer metadata. For current Amazon ranks after a growth check, use the Amazon Best Sellers API.
FAQ