Import data to Salesforce with Keboola

Problems consultants have on weekly basis – how to quickly get data into Salesforce.

A few ways everyone knows about – Import Wizard, which I try to start away as far as possible as I never know what it will do; Data Loader, which needs Java; Workbench, which can process just files up to 2MB; DataLoader.io, which is free up to 10 000 records; Talend, which you probably need to learn.

But what if you cannot use Java, if you have more than 10 000 records and you don’t really want to pay? Keboola is here to save you.

Free tier

I’ve been using Keboola for a few years, no extensively, just every than and now and I’m still surprised how „easily“ it can be extended if you have some basic developer experience.

And just lately they introduced two important changes in their payment model.

The first thing is „free tier“, where you can use up to 300 minutes per month, which is not bad at all. I tried to import a few big files into Salesforce and used a few minutes. It all boils down how effective your import can be, which goes back all the way to the best practices of disabling process builders, have clear data and all those things, including whether you can run your import in paralel or it has to be in serial mode because of locking issue. All in all – make your import as smooth as possible and it will use as little minutes as possible.

Second thing they did lately is „pay as you go“ model. Before it was always contract based and their prices started (if I’m not mistaken) around 2 000€/months. Pay as you go can put you lower, utilizing the free tier and then just pay any extras you have.

How hard is it?

After UI change they did a while ago, I feel it is super easy.

  • sign up at Keboola (no credit card required);
  • create bucket in Storage (maybe you will find out why you need to choose in or out, but it probably make sense, that things which go to Keboola are in the „in“ bucket, the others in the „out“);
Storage page
  • import your CSV data into the system, the only rule here is probably that the table name should be the same as the object name in Salesforce;
  • in the Components menu choose Writers and select Salesforce;
Writer definition
  • first thing is to define the connection things – username, password, security token, sandbox/production. Also you can choose the serial/paralel mode here, specify operation (insert, upsert, update, delete) and some others;
  • last thing – choose the tables you want to update, you can also choose only some columns to upload;
  • click the magic Run Component and hope for the best. You can monitor the job on the Jobs page;
  • done!

That’s it?

Not at all, you just scratched the top of what Keboola can do. Obviously it can extract data from Salesforce as well, it even have a Scaffolds – predefined template – which will extract the typical Sales objects from Salesforce and linked them together in Snowflake, which sits under Keboola, so you can do additional magic with it.

Obviously there are tons of other connectors – ActiveCampaign, Asana, Facebook Ad, FTP, Google AdWords, Google Analytics, LinkedIn Ads, Mailchimp, Microsoft Dynamics, Zendesk, some banks, some (especially Czech) invoicing systems.

Plus you can do transformations directly in Keboola (in Python or Snowflake directly in free tier, additional technologies supported in paid version), you can orchestrate all those things (extract data, transform, upload back).

Keboola is one big animal, which can cover most if not all of your data needs. Feel free to use it to just import your data to Salesforce as it doesn’t really have limits, at the same time you might use it way more. Go for it!

2 komentáře

    • you can use external ids to load related records, but it cannot recreate the whole structure if you give it a few tables (out of the box, with some orchestration, which is possible, doable)

Leave a Reply