Hi, I was able to succesfully setup an experiment A/B test. Unfortunately I don't see any results (see attachment 1) I tried already to setup this: https://docs.growthbook.io/guide/google-tag-manager-and-growthbook#4-tracking-via-datalayer-and-gtm And: It looks successful. My Google Analytics shows these events were fired (see attachment 2). Maybe I have an idea: When I open SQL Explorer I see data only until 20250616 (see attachment 3). On Google Big Query I see already more until 20250622 (yesterday, see attachment 4). Now I have three questions: 1. How can I make Growthbook to get all data from Big Query (manually, or better full automatic)? 2. What can I do to see my experiment results? 3. How can I set up Biq Query to read data from Google Analytics in real time (manually, or better full automatic)? Thank you very much for your help David

Hey @fierce-camera-18254! • Clicking Update on the Experiment results page will fetch the latest data from BigQuery. GrowthBook also automatically refreshes this data. You can change the frequency under Settings → Experiment auto-update frequency. • On the free version of BigQuery, there will always be a ~24 hour lag with GA 4. You'll need to upgrade to BigQuery 360 for more frequent updates. • If you're still not seeing data, metrics have minimums that need to be hit before they are shown. If you have a small amount of traffic, then it could be that these minimums aren't hit yet, which might prevent results from being shown. You can see/configure minimums on per metric basis but also under the Metrics & Data page in Settings. • See also: https://docs.growthbook.io/kb/google-analytics/google-big-query-bq-troubleshooting

Hey Ryan, and thank you for your valuable tipps and idas! • I tried already to press on "Update", that was working some 6-8 seconds, but still no results visible. The "Experiment auto-update frequency" was already set to 6 hours (I guess it is the default). Can I change that to 1 hour or 1 minute? What would be the implication? • Ah, I see. That invokes immediately two questions: a) Can I create a manual "Update now", and if so how can I do that? b) How can I do that upgrade to "BigQuery 360"? • Ah, and I still see no data. For that experimant yesterday I ran ads to that page, leading to 383 people clicking on my ad and going to my landing page. I guess that's more than enough. Now I'm checking "Minimum Metric Total" in the settings as suggested, and it shows 150 "total metric value required in an experiment variation before showing results". I changed that to 50 now and checked my experiment again. Result (after pressing on "Update" again) is still: "No data yet. Make sure your Experiment is tracking properly." • I also checked your recommendet link "https://docs.growthbook.io/kb/google-analytics/google-big-query-bq-troubleshooting" but it doesn't contain anything useful for my situation (or at least I didn't see it as helpful). Do you have another idea, how to make my experiment results visible?

Thanks for running through all those steps. Let's check two things: 1. For the goal metric in question, go to Metrics & Data → Metrics → Your Metric. Under Metric Analysis, does it show data? It'd look like the first screenshot below. 2. Go to Data Sources and choose your BQ data source. Under Experiment Assignment Queries, click the ⋮ (three vertical dots) and choose Edit Query → Customize SQL and Click Test Query. Is data returned like in the second screenshot?

Thank you for your help again, Ryan 1. It shows data, but it seems to be very old (see screenshot 1), last data is on 19th of June (but only a little bit). 2. Under Experiment Assignment Queries I found two possibilities "Anonymous Visitors" and "Logged-in Users", I choose anonymous because my traffic is not logged in into my website. Here I found outdated data like in Screenshot 2 Does that help you a little bit?

If you click on the Rendered SQL, can you share a screenshot of that?

Yes of course, but I don't find an "Rendered SQL". My Sql command is this one, do you mean that?: SELECT user_pseudo_id as anonymous_id, TIMESTAMP_MICROS(event_timestamp) as timestamp, experiment_id_param.value.string_value AS experiment_id, variation_id_param.value.int_value AS variation_id, geo.country as country, traffic_source.source as source, traffic_source.medium as medium, device.category as device, device.web_info.browser as browser, device.operating_system as os FROM

total-cascade-452322-m7

.

analytics_398464933

.

events_*

, UNNEST(event_params) AS experiment_id_param, UNNEST(event_params) AS variation_id_param WHERE ((_TABLE_SUFFIX BETWEEN '{{date startDateISO "yyyyMMdd"}}' AND '{{date endDateISO "yyyyMMdd"}}') OR (_TABLE_SUFFIX BETWEEN 'intraday_{{date startDateISO "yyyyMMdd"}}' AND 'intraday_{{date endDateISO "yyyyMMdd"}}')) AND event_name = 'experiment_viewed' AND experiment_id_param.key = 'experiment_id' AND variation_id_param.key = 'variation_id' AND user_pseudo_id is not null

I'm not seeing anything wrong here. Going back over the screenshots, the only thing I noticed was that there were several experiment IDs. For the experiment that's not showing metrics, are you able to check in BigQuery that there are a significant number of exposure events?

I'm glad to hear that from you, that you don't see anything wrong 🙂 I don't know how to do that, but maybe this attachment is interesting, because it only shows very little event rows (9 in total). Does this help?

Yep! I think the issue is that there isn't enough data to show yet, which is why you're getting that error. Once enough time has passed, you should see some results 🤞

Sooooooo, now there is much more data in Big Query (see screenshot, 200 rows already)! But: Still no data in my experiment. But I see in the screenshot (on the left side) that there is now two analytics_NUMBERS - and I bet Growthbook doesn't now (yet) that it should look to the new data. Unfortunately I can't find how to change that... Do you have an idea?

Are you able to share the page where this is running?

Yes of course. Here is the realtime page in Analytics

sorry, I meant the webpage—the actual page where the experiment is running.

Ah, haha. This is here: https://davidkirschey.de/lp-depressionen-behandeln-ohne-medikamente-ohne-arzt-ohne-tabletten/

I don't see any experiment active on that page?

Ah, that's interesting. But it is active... (see attachment):

It was because of the cookie consent widget. Until I consented, the experiment doesn't load. That could explain the low traffic count

Sure, that's a godd idea. But as you consented, I should have at least ONE experiment result - but there is no (see screenshot). What can we do now? Without experiment result, GrowthBook is nearly useless for me, you know? Do you have another idea, how I can get my experiment results visible? I'm so glad, that you help me, Ryan. Without you I would be lost totally!

When I initially went to the site, I did see a tracking callback fire for the experiment, so you should see that in Google Analytics. (I no longer see the experiment on the page, tho). If you only have a handful of users, though, there won't be enough events to run the test. Can you see now in GA how many

experiment_viewed

events there are?

Haha, just today I stopped that experiment because I started to send fresh traffic on that page, but I modified the page, so the old experiment didn't fit make sense more. Actually we have 300 users daily on the page. Wait, let me start a fresh experiment... Okay, I duplicated the stopped experiment, made a simple change with visual editor... And now the new experiment is live. Now comes the important question: How can I see how many "experiment_viewed" events are in GA?

See it on there now! So, it should show up in Google Analytics under Ereignisname (I think it is in German). So, you'll see like page_view next to experiment_viewed

Yes, I see them now, too (see screenshot)

Nice. In GrowthBook, then, you'll need to hit the threshold (around 150 events per variation) to start seeing results.

Great. Why do we need such a huge amount of events to see results?

Go to your GA/BQ page under Data Sources. Go to the Experiment Assignment Queries. Choose the one you're using, click the three vertical dots and click Edit Query. Then, click Customize SQL and Test Query. If everything is configured correctly, then rows of data should be returned in Results. If not, then something in that query needs to be updated.