Do you know a tool to create customer facing analytics for our SaaS?
Giorgio Groß
16 replies
Hey, I'm founding a B2B startup we're currently integrating customer facing analytics into our product, e.g. where our customers can see the performance of our tool. Ideally I want to be able to track analytics events with a single line of code (from our customer's websites) and run queries against the data set with the second line (from our customer dashboards).
What I found so far (but didn't perfectly meet our needs): Verbdata, keen, Datadog, Segment, aiven, Rilldata
I'm this close (you can't see me but I'm holding my thumb and index finger really close together right now) to implement our own solution. In case I do that, my second question would be: Would you be interested in me sharing my approach or use that as a service?
Replies
Pablo Fatas@pablo_fatas
At SigmaOS we use Mixpanel. While we have some greviances with it, I think it would be great for what you are looking for.
Share
@giorgio_gross Hmm thats fair. It can get quite expensive at high usage. We have also had to cut down on some tracking to make it more reasonanble now that have credits are expiring. Good luck on finding something better and let me know if you do. definitely feel like there is a lack of tools for this kind of things
@pablo_fatas Thanks for the suggestion! I think the problem with Mixpanel would be that we need to create custom dashboards for our own customers from the data. As I understand Mixpanel is mainly intended for cases where we would want to track events from our own website. And it's quite expensive for us (we'll track more than 10K MTUs per customer)
If I understand well you want to track events on the customer website, not only on your own website. Is that it?
@fabian_maume Yes exactly, and we want to create some charts for that customer with the events we log there
@giorgio_gross Then I would recommend to create a google tag manager template. It will make it quite easy for your user to deploy it on their website: https://developers.google.com/ta...
@fabian_maume That is indeed something I didn't even look at! I'll check it out and see how it could work :) Thanks man!
I build LogHarvestor. It's like DataDog but for startups, and it has a free plan. It's got a good API and currently integrations for GoLang, NodeJS, & Python. It processes all events/logs as JSON, so it's super flexible. I don't have a ton of users for it right now, so if you wanted to join up I could build features that you specifically want.
If you're interested let me know:)
Here is a link to the site: https://www.logharvestor.com
Oh, and if you want to play around on the demo account here is the link for that:
https://app.logharvestor.com/demo
@solomon_bush Hey that looks interesting, currently checking it out. How do you perform if there is ~1k simultaneous logs arriving? And can I run queries against your logs through the API? How would they perform if there are millions of entries in the database?
@giorgio_gross Terrific and great questions!
So LogHarvestor uses node.js servers to process all the log requests. This is because it's asynchronous by default. My servers can handle ~40k req per second, and I have at a minimum 8 running in a cluster. I also have auto-scaling enabled, so it will be able to handle whatever you can throw at it.
LogHarvestor's databases are also cloud hosted, highly-available, and have auto-scaling enabled. The databases can also scale horizontally, so performance doesn't take a hit when scaling up the amount of logs.
I am still working on refining the aggregations and queries to improve performance, but right now if you have several million logs and want to run a full-text search, the response time is less than 3 seconds.
Currently we support full text search via the API, which can be filtered, limited and sorted. I need to update the API documentation for this. But for an example in the demo-app, see this link, while logged in as the Demo Account: https://app.logharvestor.com/log...
Also, v.1.0.0 of LogHarvestor will be launching this summer. It's going to have loads of new features and expanded query features.
@solomon_bush that sounds cool, I'll dive further in to see how it fits in! For a production usage it would be good to know how resilient it is e.g. how strong is your tech stack, what it's built on etc. Have you any larger reference customers already? I know how hard that question can be, I'm also building a SaaS and in the beginning it's kind of hard to get the first big customers to trust haha, but would be interesting to see how others use your service
@giorgio_gross That sounds great! Would you be open to a phone call or zoom call some time? I would love to find out more about what you're building and answer any questions like this you have.