Data Science & Analysis

Power our next data science project by taking advantage of Narrator’s webhooks and views
A/B TestingAd Hoc QuestionsData Science ProjectsMarketing AttributionRecommendation EngineReuse Analytical Approaches
Use Cases

For Data Science Projects

A/B Testing

Evaluate the impact of each treatment in Narrator
Icon Arrow Right

Ad Hoc Questions

Quickly answer stakeholder questions without needing data engineering's time
Icon Arrow Right

Data Science Projects

Power your next data science project by taking advantage of Narrator's webhooks and views
Icon Arrow Right

Marketing Attribution

Build datasets to quickly compare and implement attribution models
Icon Arrow Right

Recommendation Engine

Power your recommendation engine with Narrator's webhooks
Icon Arrow Right

Reuse Analytical Approaches

Narrator makes it easy to re-use analytical approaches again and again for different questions
Icon Arrow Right

A/B Testing

Evaluate the impact of each treatment in Narrator

A/B testing platforms are expensive and they are often limited to the data that is captured in that platform (often only web conversions). With Narrator, evaluating an A/B test is easy. Log the variants each customer has seen and build detailed analyses to understand how each variant influences your bottom line. Bonus: That analysis can be reused for future A/B tests too.

Benefits

  • Understand the downstream impact of any experient
  • Apply a consistent statistical approach and read-out for each experiment
  • Build health-checks into your A/B testing evaluation
  • Track the impact over time to see if the influence has been consistent
  • Keep a record of all experiments, their outcomes, and the business context in one place

Applications

Readout of Results

Create a Narrative (live-analysis) to give a read-out of results in a story like format. Ensure that the interpretation is always up-to-date by including dynamic logic in the text (ex. if needed sample size is met then interpret the results else wait and see).

Split Traffic

Capture User Data

Analyze Results

Ad Hoc Questions

Quickly answer stakeholder questions without needing data engineering's time

Narrator makes answering ad hoc questions a pleasure. Stakeholders are always asking for "quick" questions but those questions require Data Engineering to model data and BI to build a dashboard. With Narrator those questions can be answered in minutes. By having all the data in one place and ready to use, there no question that your stakeholder can ask that you cannot answer.

Benefits

  • All the data is available for you
  • Never worry about source of truth because you are always using it
  • Missing Foreign keys are no problem because anything can be related thanks to Narrator's innovation

Applications

Exploring Data

In just a few minutes, you can assemble a dataset that bridges multiple data sources and quickly visualize the trends over time. This makes it easy to explore trends and dive deeper into your analysis.

Getting KPIs

Calculate company KPIs quickly. When a follow up question comes and you need to slice your metric by another feature, Narrator makes that as easy as a couple of clicks.

Ad Hoc Question

Build Datasets

Quickly Give Answers

Data Science Projects

Power your next data science project by taking advantage of Narrator's webhooks and views

Data science is hard enough without having to deal with getting the data you need. With Narrator, you can create the data you need, save it as a view and pull it in your Jupyter notebook in seconds. As you need additional columns or new data, you can simply update the dataset and go back to your algorithm.

Benefits

  • Your training datasets will always use business logic that's adopted by the rest of the company
  • Deploy your algorithm quickly when needed
  • Keep your notebook clean without many JOINs
  • Easy to update your dataset

Applications

Kalman Filter

Use Kalman filters to predict churn by pulling time series data into your notebook, using the aggregations as training data. This allows for a live training set with all the customer's behavior data available, trusted, cleaned, and ready to use.

Logistic Regression

Creating a supervised dataset in Narrator is easy. Leverage the default conversion columns ("Did activity"). This will create a table with many 1/0 flags, which is ideal for logistic regression.

Bonus: Reuse your algorithm 👏

It is easy to swap out activities then reuse your algorithm. Your model to predict churn can be re-trained and used to predict signups with a couple of updates to a dataset.

DS Notebook

Create a View

Narrator Dataset

Marketing Attribution

Build datasets to quickly compare and implement attribution models

Marketing attribution is crucial to understanding the effectiveness of your marketing spend. Unfortunately, ad platforms like Facebook and Google give a siloed view of performance and over ROI is overstated because the platform is not integrated with other systems. In Narrator, you can create marketing attribution models quickly and accurately. First touch, last touch, any touch, linear, etc. can be created with a dataset in minutes.

Benefits

  • Uses the data from your company's source of truth
  • Expand your attribution model to account for edge cases
  • Explore various attribution models in minutes
  • Quantify and debug unattributed customers using customer journey

Applications

Operational Spend

Set up operational dashboards for your marketing team so they can make spend decisions in real-time.

Marketing Spend Strategy

Evaluate which platforms have the best returns, which are best at driving new customers and which are best for remarketing.

Narrator Dataset

Materialized View

Marketing Results

Recommendation Engine

Power your recommendation engine with Narrator's webhooks

Recommendation engines are crucial to your business but can take months to implement. With Narrator, you can set up a recommendation engine in under a day

With Narrator, you can assemble a dataset that ranks the feature you want to recommend by your desired outcome metric. Send the ranked feature to your product or website via a webhook integration. That's it!

Learn more about webhooks

Benefits

  • All reporting uses a single source of truth
  • BI platforms are fast because JOINS are not in the BI tool
  • Trace metrics to their source in two clicks (open the dataset and see the activities)
  • Complex logic is kept in managed by you in your data system, not in your BI tool
  • All data model updates are done in one place

Applications

Rank Search Results

Rank search results for each keyword based on likelihood to purchase

Recommend Entry Products

Maximize customer LTV by dynamically surfacing product recommendations for first-time customers

Surface Cross-Sell Opportunities

Recommend the the next best product for a customer based on their prior transaction history and trends from other customers like them.

Narrator Dataset

Webhook Integration

Website / Product

Reuse Analytical Approaches

Narrator makes it easy to re-use analytical approaches again and again for different questions

Consultants are always recreating common analyses for each new client, but the approach is often very similar. With Narrator you can create an analysis once and reuse it again and again. The standard data model makes it easy to copy and paste analyses that read like they were purposefully built.

Benefits

  • Build an analysis once and reuse it again and again
  • Data transformations don't need to be re-created
  • Faster delivery of each analysis

Applications

Data Consulting

For the last 3 years, Narrator provided quality analyses for customers. We were able to make it affordable because a single Narrator analyst could support eight companies as their data team. The magic was the fact that most of our analysis were templated but read as if they were purposefully-built.

Learn more about re-purposing datasets

Narrator Analyses

Swap Activities

New Analyses

Narrator © 2020