Subscribe
Sign in
Narrator — All your data, one table. A new standard for data modeling.
Narrator replaces a star schema with one 11-column data model. Their new self-serve product transforms your company’s data into a single 11-column data model and allows you to generate any table for BI, reporting, and analysis using just that model.
42
Replies
Garry Tan
Hunter
Narrator is the holy grail when it comes to understanding your user behavior. Any founder or data scientist who has tried to join data across lots of different sources (e.g. your User table, web analytics, email analytics, data warehouse) will run into incredible complexity: whether it’s having to use a star schema or lots of weird inner joins and subqueries. Narrator lets you skip all of it and have a single view that lets you answer deeper questions with way more clarity than you’ve ever had before. Now Narrator is ready for you to try for free with your own data. Once you try it, you’ll find you can answer questions far faster and you won’t want to go back.
Share
I’d like to look at my pipeline side by side with my Stitch / DBT powered pipeline. Are there docs to help? And also any advice on how to run a comparison would be great.
@scottswitzer Great question!
Narrator can run side by side to your DBT transformations.
What I would recommend is to connect your warehouse to Narrator and go through one of our Choose Your Own Adventure tutorials: https://docs.narrator.ai/docs/ge...
Then choose a question that you would have to write in a custom DBT query for and assemble that same table in Dataset.
What you will see is that that table will be :
- A couple of clicks instead of a lot of SQL
- Way faster
- Never duplicates or looses rows (EVER)
- Can be assembled even when there is no foreign key (which is common when bridging datasources)
Narrator is DIFFERENT but once you try it, I think you will join us in not being able to go back to traditional data modeling.
I brought Narrator into my company, Candid, and as their very first customer I have watched them evolve this product over 3 years. Impressive team and product. Hit the buy button.
So dope!! Congrats Ahmed and team. Definitely see a lotta exciting applications. Hit me up via email.
Would love to run a few ideas by you on ways we could collab!
@danieloduffy HAHAHA we needed to share our learning to enable other people to not suffer with data like I did.
We've spent weeks/months fighting with data instead of learning from it because something like this didn't exist. Never again... will rely on Narrator from day 0 going forward.
@nick_alexander Thank you so much! I have been there and know the pain you're experiencing. Numbers not matching, looking at how a materialized view is defined to see another MV and then giving up, and finally spending weeks working with a table to learn that there is a v2 with the right data.
I am very proud that this is no more and am excited for you to experience more of the power of data without all the fighting!
Thank you @garrytan for this hunt!
Hi Product Hunt, I’m the CEO and co-founder of Narrator. We’re excited to finally share our new self-serve product with the Product Hunt community!
I, and most other analysts, have experienced the challenges of data–numbers not matching, spending hours trying to find the right data, and dealing with endless maintenance of dependencies.
We found that most of these issues were because of the star schema approach, which cannot handle the current rate at which startups are changing. As new questions arise (which they always do) the number of data models increases, thus it becomes impossible for your data team to maintain a single source of truth, consistent metrics, and quality analyses in a timely manner.
Narrator’s new self-serve product is both an engine to transform your company’s data into a single 11-column data model, and a platform to generate any table for BI, reporting, and analysis using that data model. As it’s only one table it’s a literal single source of truth. Therefore, there is no web of dependencies, and the nightmare of making changes to BI tables (adding new columns, updating logic, etc.) can be done in a few clicks without adding any complexity to the definition.
If this sounds unbelievable, leave a comment with literally any data question and I will translate it to the Narrator world. I have been doing this for years now and am excited to see what questions you all come up with.
Looking forward to diving into data modeling and showcasing the power of a time-series table in the comments below!
Check out some awesome use-cases from our customers: https://docs.narrator.ai/page/us...
@ahmed_elsamadisi @garrytan we are looking for interesting data startups to participate in Cloud Data Summit. Would love to chat about getting Narrator involved
@ahmed_elsamadisi @garrytan @eric_axelrod that sounds great! Wanna reach out over DM and we'll get it set up (contact@narrator will go right to me)
Been using narrator for a while and it’s been so helpful! Can’t wait to see what comes next
Few ideas are so brilliant as this one I believe, congrats. Do you also plan for some sort of 1 click analysis I can run on my data?
@vito_margiotta Yes! That will be coming soon. The Narrative (actionable analysis in a story format) can actually be converted to a template. Because Narrator uses a standardized data model across companies, that template can literally be run by anyone. In the near future, Narrator will release about 100 templates so everyone can get access to rich analysis that we have built.
In the future we imagine a store where any Narrator user can create, sell or buy someone else Narratives in a store.
That is a future where analysis is democratized!
I've been using Narrator for over a year now and I can confidently say that I'll never go back to writing queries or building custom data models. Star schema just isn't worth the investment in time/maintenance - I rather spend my time on analysis and insights!
This looks super useful - how would I use this to analyze my e-commerce users' data?
@jennifer_rose_goldman Awesome question.
For e-commerce data you can model your data as a customer doing thing in time: visited site, viewed product, completed order, shipped, submitted ticket.
With dataset, you can answer any of the following questions with a couple of clicks.
- What ad source gets the best customers to optimize marketing spend?
- What is the best product to show on the home page to increase conversions?
- What is the best email to get someone to repeat purchases?
- What product gets the most ticket complaints right after being delivered?
If you can describe a question about your customer then Narrator can assemble that dataset.
Do you have some questions you are trying to answer? I can help map them to Narrator questions.
@jennifer_rose_goldman @ahmed_elsamadisi Sounds really interesting, would I be able to calculate something like "New Users" or "Unique Customers" over some time period?
@jennifer_rose_goldman @maxcobert yes new users and unique customer can be done in a couple of clicks.
You can go way further than that by diving into behavior.
@jennifer_rose_goldman @maxcobert yes. Happy to show you how this is a couple of clicks in Narrator. We now have demos scheduling on our site, so feel free to schedule time and we can answer this and any other question you have
@harsha_rao good question!
Narrator breaks down your business into customer activities. Each row represents a customer doing a single action in time and we only need 11 columns to represent those activities. Think timestamp, customer identifier, activity name, a couple features, and some identity resolution columns and that's it!
We have concrete sql transformation examples here:
https://www.narrator.ai/how-it-w...
And a deeper breakdown of the 11 columns themselves here:
https://docs.narrator.ai/docs/ac...
Hi @mattjstar! Super interesting! Does this mean that you’ve found that the majority of data questions any business can have about their users can quickly be answered by running an analysis off a table of just 11 columns? If so, does that imply you’d recommend any org to create an activity stream of these 11 columns to speed up their analyses? Or does your platform use this data model in a special way that traditional tools would not?
@mattjstar @harsha_rao Yes! Everything happens in time, we model those into activities in time. You describe a question by talking about a customer's actions in time, and that's the same exact way we assemble the data.
The challenge with companies creating their own activity stream is that it requires a lot of work to get it to be accurate (updates a bit differently, identity resolution, optimization).
Then actually querying the activity stream is extremely hard since there are no foreign keys. Narrator reinvented joins to get a single activity stream to work.
If a company solves all the problems of setting up and activity stream and build a powerful UI that can assemble any table on top of the activity stream without joins then that company would be Narrator.
@tarush_aggarwal1 Thanks. As a leader of a Data Engineering team, do you find that simplifying modeling opens up your team to focus on analysis and other parts of the business?
