What is a Fused Warehouse?
A fused warehouse consists of records that have been merged and normalized. So if your company uses multiple SaaS applications — Salesforce for leads, Marketo for marketing automation, Zendesk for support — one customer might have a separate Contact record for all three channels.
If you automatically matched these three Contact records, then merged and normalized them, you’d have yourself a fused record, free of conflicts and duplicates, which you could instantly query and visualize with your tools like Tableau, Microsoft PowerBI, Amazon QuickSight, MetaBase and more. In your SQL output, you’d see zero repeats and consistent formats everywhere. For more, read our other post "Why Fuse Your Data?".
How Does Bedrock Data Fuse Records?
At a high level, fused records get made in two steps.
1. Form relationships between records in multiple systems.
2. Combine records from multiple systems into a single entity that is easy to access.
Fusion’s graph database takes care of this first step. By connecting data points, or nodes, it’s possible to see relationships between records in discrete SaaS applications. This is because nodes contain attributes (key: value pairs, e.g. Contact: John Doe). And those attributes with a common identifier (e.g. Email: firstname.lastname@example.org) allows Fusion to map relationships, or edges, between systems. Each relationship between nodes has one or more properties (purchases, ratings, time, distances, strengths, etc.) attached to it. Fields are affiliated with a data type (e.g., integer, float, decimal, string) and parsed as an identifying field (like an email address) or non-identifying field (like a product), so relationships properties are mapped in the graph.
By mapping the identifying fields, Fusion can tie one field (like a Company ID) to another (like an email address). Mapping external APIs allows us to transform the identifying and relationship fields by applying standards to them.
Data rich in in relationship properties is much more valuable than if the records were siloed. Such data can feed back into the original applications to make them more complete, accurate, and up-to-date.
To create a database of fused records, these multiple records must now merge into one document, during which process data are “standardized”. For example, if the email email@example.com is recorded in my Salesforce and JohnDoe@acmecorp.com is recorded in my HubSpot, the fused record would standardize the capitalization to a more consistent lower-case format.
A fused warehouse blends your fused records together, giving you access to the clean and normalized tabular data of a relational database, devoid of duplicates and format issues, as well as the relationships between objects in systems.
Then, once standards are applied, these fused records get written out to the external systems which fed them in the first place.
How Do I Access & Use My Fused Data?
Now you can start querying the SQL dataset which contains all your data sources. By accessing a fused warehouse you can generate complex reports using any analytics tool of choice — Looker, Domo, Chartio, whatever you please. A task vastly easier than had the data still remained siloed within their original SaaS applications.
Let’s see some examples.
Say you use GoToMeeting, HelpScout, and HubSpot. In a tool like MySQL Workbench, you could run a query that links fused records from the Company_Tickets_Links table, returning the status of the ticket, case numbers, normalized dates, and more.
Say your support data is in HelpScout and your marketing automation data in HubSpot. How can you link tickets to certain companies, matching support cases to actual accounts, when HubSpot uses a Company object, and HelpScout a Contact object?
Without Fusion, you’d have to look at multiple columns in the warehouse to join. In HubSpot’s data, the company might also be Acme Inc., in HelpScout’s it could be acme. Just to answer this simple question, you’d have to do multiple matches, look at multiple columns.
With Fusion, all these matches happen natively. You simply link HelpScout’s Tickets table to HubSpot’s Contacts table — et voilà, all tickets are linked to company accounts. To visualize, you could use Tableau or any other BI tool.
This simple linking opens up the 360-degree view of your customer so vital for onboarding customers. But this you could do in an endless number of ways. In another simple example, here we get a count of which pipeline stage all of our customers are in by fusing data from multiple SaaS applications, this time with MetaBase:
How does fused data compare to data from other ETL tools?
Unlike other ETL tools, which make you clean, combine, and warehouse the data yourself, Fusion extracts your data and automatically provides both the fused records and access to the warehouse storing the schema of fused records, both of which may feed your analytics, BI or reporting tools.
Back to the Fuser: A.I. & Machine Learning
Thanks to the rise in A.I. and machine learning, in the future fused records could spark actions — automated changes to records in real-time, webhooks and workflows for actions such as sending a notification to an account manager, or contact to a customer via social media.
There are a lot of great reasons why you should fuse your data. Today, getting started with your trusted data source is simple. Here’s how it works:
1. Connect your data sources. Extract your data from SaaS applications such as Marketo, HubSpot, Salesforce and NetSuite. In just a few clicks.
2. Fuse your data. Automate data matching, de-duplicating, resolving data conflicts and modeling object relationships for more real-time reporting and dashboards.
3. Warehouse your data instantly. Automatically access your warehouse via a cloud data warehouse.
4. Feed analytics & BI tools. Let your cloud data warehouse feed SQL input into tools you are using for visibility and analytics including Tableau, Microsoft PowerBI, Amazon QuickSight, MetaBase, and more.