All the documentation I read is for creating a custom application from a local environment. I don’t see anywhere a document explaining in detail how and where to place the code for sending conversion events to the Algolia dashboard. I have followed along to the existing doc and have been successful in replicating an application that links my shopify index and pulls the products upon searching. I have used the chrome extension “Algolia events validator” and it all works perfectly. I have pasted the code snippets all over the files in the Shopify environment but nothing works for me. I was charged a premium for an upgrade in order to use the recommend features (“Frequently bought together” ,“Related products”) and have not been able to train the models due to the fact that I cannot send conversion events. The sales rep sold me an upgraded package based on conversions and I didn’t actually have any conversions sent. At the time I was new to Algolia and didn’t realize that I had 0 conversions. I have contacted support and they always refer me to Algolia docs that say nothing about how the Shopify file system works. Can anyone help? Thx
sorry to hear that you have issues with getting conversion events from your Shopify app.
Are you using the default widgets provided by the extension? If so, did you follow along Click and Conversion Analytics | Shopify | Algolia and are still facing issues?
Also note that Recommend has a beta feature that allows sending historical events; you may want to check Uploading existing events via CSV | Algolia in case that could be a fit for your needs.
I was able to send the events. I did enable the click analytics and followed along to the documentation. The code in the docs had a line that needed to be changed in order for it to work, so no amount of following along as the docs say would’ve helped anyone.
Also, I have tried to send historical events and the csv format is never correct and algolia support just refuses to send a complete csv example of how the data should be structured.
The suggestions they’ve made regarding this has never worked and I always receive an error message saying the file is too big or the format is incorrect.
I am now receiving the error massage enclosed and it’s very frustrating that I’ve been following along to instructions, documentation and support suggestions that now, almost 3 months later have lead me nowhere.
Beginning to think that Algolia is just unclear purposefully to bully customers into purchasing an integration package or something. This shouldn’t be a problem the way it’s been.
I can understand your frustration. Some of those features are still pretty new and we need to improve our documentation. If by any means you remember the misleading line in the Shopify documentation, that would be helpful to avoid future people to have the same issue.
For the csv upload, from what I can see, any csv, with
, as separator and the following list of columns (in any order) should work:
"userToken", "timestamp", "eventType", "eventName", "objectID" and any column outside of this list would be ignored.
Then each value should follow the format specified in the Insights API documentation: Insights REST API API Reference | Algolia.
We clearly need to improve the UX to help debugability on this feature. If you believe your CSV should work and the CSV parser might not work as expected, I’ll check with our engineering team working on the feature.
Yea I’ve spent months doing this. No amount of editing the csv file ever worked. At one point your support team admitted it was an issue on their end and to try in different environments/browsers, nothing worked. I tried at first collecting events and could not do so. Turns out, documentations was not correct. So naturally, I went with trying to put together a doc based on actual events and product IDs within the last thirty days(during that time) to have relevant data. That was not easy nor fast. I was charged a premium based on a number of conversions that I was never actually collecting. Very little help from support. Communication with support takes days before a response, usually. Really disappointed. My team is over it.