Blend reaches terminal velocity with Fivetran

The fintech startup uses ELT and Reverse ETL to deliver massive business value


“The value of Fivetran isn't any one individual connector – the value is being able to pull in Salesforce, Marketo, Asana, NetSuite and Lever, and blend the data from historically separate departments together for analysis. With Hightouch we can then push it out to make sure that everyone's looking at the same metrics.” –
William Tsu, Customer Success Operations, Blend

  • Blend’s Finance department is able to close out service team books four days earlier, cutting the team’s financial reporting time in half
  • The Operations team merges data sources together to provide powerful, business-wide data analysis
  • Hightouch enables alignment between teams and tools, driving better alignment and impacting Blend’s bottom line

ELT: Fivetran

Reverse ETL: Hightouch

Warehouse: Amazon Redshift

Connectors: Salesforce, Marketo, Asana, NetSuite, Lever

At Blend, velocity is vital. Founded in 2012, the fintech startup has undergone a rise to prominence in the banking software space, announcing their IPO in July of 2021. In the last year alone, Blend enabled financial services firms to process nearly $1.4 trillion in loan applications, handling an average of more than $5 billion in transactions every day. On the face of it, Blend’s vision is simple: to re-architect banking software around the consumer. Customers move fast, and increasingly expect their bank to provide a customer experience in line with the service offered by other disruptive, digital-first applications and businesses. But delivering that experience and agility requires massive internal coordination – something that is central to the roles of Ryan Dunlap and William Tsu.

Ryan and William work on the Customer Success Operations team at Blend. Their team’s purpose is to build out large, innovative business systems to enable the business to move quickly: whether that involves Salesforce, SQL, Databases or Asana.

“Our general principle is optimizing for velocity,” says William. “So, how can we get to market quickly? Being able to pull that data into our data warehouse, turn them into insights, and being able to feed that back into our key stakeholders so that they can be informed about the business is probably the biggest value add that we do.”

The Challenge


From its inception, Blend had adopted a modern data stack approach, with Redshift at the core of its analytics operations. Getting data into and out of the data warehouse, however, was another challenge entirely.  

“We did not use third-party vendors for that. Everything was in-house, and it was all screwed together via Zapier, Airflow operators, Airflow DAGs, writing up our own code to hit up the various endpoints, in particular the Salesforce endpoints,” says William.

The convoluted approach to data ingestion was costing the team time and money. Pulling just a single column from Salesforce or changing a field could take the team weeks, limiting access to time-critical data. Without the ability to prototype and rapidly iterate, the team had to release straight to production to be able to test their solutions – creating more headaches for the operations team. And with rapid expansion came the onset of new tools to manage workflows and processes, like Asana, Marketo, or Lever. Each of those tools also required data to be synced inside them to be more effective.  

For Blend – a business built on agility – any level of slowdown was unacceptable. With the data engineering team’s limited bandwidth, they did not have the capacity to expand their scope to maintain a rapidly expanding list of SaaS platforms. William and Ryan faced a decision: commit to in-house tooling, or look for external providers.

“Our in-house tooling just didn't scale,” says William. “We were hitting enough data where it took a very long time to load into our system and push back out, or just was not working any more.”

The Solution


William and Ryan laid out clear criteria for their decision making process when assessing ELT and Reverse ETL solutions:

  • Ease of use. “How easy is it to use, in particular in the context of velocity?”
  • Connector support. “What connectors do these products support? If they don't support the tools that we're using, then they're not really what we're looking for.”
  • Infosec requirements. “Very important because we're a financial technology company. For any tool that we use, we need to be able to store the data within our own country. You can have as many bells and whistles as you want, but if you don't support that paradigm, we just can't do it, full stop.”
  • Alerting. “When things are down, do I have the ability to troubleshoot to know why things went down? Is there a team that supports us to get it back up? Alerting is definitely key. Having an audit trail when something goes awry is pretty important.”

From an operations perspective, Ryan was looking for a tool that was self-serve and would allow the team to deploy rapidly, with little to no dependency on data engineering.

“We've found to have more success with a purpose-built tool, versus something that does everything ‘okay’. There's certain things you just want to do well and not have to worry about it, on the ELT side with Fivetran and then on the Reverse ETL with Hightouch,” says Ryan.

Fivetran was a clear frontrunner for Blend in handling the extract, load and transformation process. When it came to Reverse ETL, Hightouch’s impact was immediate, and transformative. The Hightouch debugger and Slack integration is now used by the Blend team to understand when the Reverse ETL process has failed, enabling a greater degree of trust in data accuracy across the board.

“Entire batches used to fail outright, and we ended up with obnoxious holes in the data. We'd have to go rerun to clean it up. Now a record fails on Hightouch for an object and we know immediately,” says William. “One of the things we love about Fivetran is the schema drift handling in the database, with no reliance on engineering. I don't have to file a ticket just to get a column added.”

One area of the business that has benefited from access to reliable data is the finance team. Before Fivetran and Hightouch, mandatory revenue calculations would take up to seven days.

“Prior to Hightouch and Fivetran, we didn’t have Asana or time tracking data. We actually have metrics now that we can look at thanks to Fivetran for the data extraction, and then we can ensure consistency of that across teams thanks to Hightouch,” says William.

Thanks to Fivetran and Hightouch’s Asana connectors, Blend’s finance team is now able to holistically monitor and enforce time entry, understand project timelines, and ultimately close out  services team financial reporting in three business days, where it previously took up to seven. The Professional Services team – responsible for integrating Blend’s solution with banking customers – is now equipped to better forecast capacity, allocate resources, and make staffing decisions based on historical data.

Blend operationalizes their Asana data using Hightouch’s Slack integration. Members of the Professional Services team are automatically reminded via Slack of upcoming deadlines (based on the task deadlines in Asana), and potentially erroneous deadlines are also flagged. Since senior management isn’t checking Asana regularly, major project milestones are automatically communicated in Slack on completion of key Asana tasks, which provides visibility to leadership and helps them know when to get involved.

“Trivial things used to take us the better part of four days to ship. We're doing most stuff the same day now. It's just given us infinitely more flexibility to support and take on new projects where we don't feel like we're constrained by the tools any more,” says Ryan.

The process of creating net-new pipelines used to take Blend’s engineering team up to a month; the Blend Operations team estimates that the self-serve process now takes the Enablement team just under three working days, and time to modify existing pipelines has gone from an estimated two days to just a minute. Before Hightouch, Blend didn’t have the ability to create new pipelines out of the business – but modifying existing Hightouch pipelines now takes a few days.

The Outcome

“The value of Fivetran isn't any one individual connector – the value is being able to pull in Salesforce, Marketo, Asana, NetSuite and Lever, and blend the data from historically separate departments together for analysis. With Hightouch we can then push it out to make sure that everyone's looking at the same metrics,” says William.

Historically, Asana data was siloed within the platform. If teams wanted to connect project timeline data to Salesforce, they would be forced to manually copy fields and values between the two. Rather than being forced to context switch between tools multiple times throughout the day, data is now shared between the platforms. “Our Professional Services Team lives in Asana, not Salesforce. To provide them with the data they need, we pull that Salesforce data into our warehouse via Fivetran, and Hightouch pushes from the warehouse straight into Asana,” says William.


By connecting contract data from Salesforce with hour logs from Asana, the Blend team can accurately assess the holistic ROI of projects – understanding where the business over or under-serviced an account, or even predict shortfalls in staffing before they happen.

The ability to connect product data to Salesforce data using Hightouch is also vital in ensuring  visibility between teams, and establishing the warehouse as the single source of truth that all teams can access in their tools of choice. If a customer is trialling a new feature within the platform, that information can now be persisted to Salesforce, giving sales and customer success teams a clearer picture of account health and activity. This also encourages Sales and Customer Success to work together, as a sales rep can’t close an opportunity until they actually see the product usage flag updated in Salesforce.

“Before Fivetran or Hightouch, you would just have to take their word for it. Now we can pull product data, compare it with what's in Salesforce, we can go back and say, ‘Oh, actually, it doesn't look like it's turned on,’" says Ryan.

By driving adoption of the product, the enablement team is able to efficiently increase the business’ bottom line.

“One of the things we can really understand is where in the project we are running into delays – where in the past we just didn't have any data to be able to look at that at all. The quicker that's done, the quicker revenue flows in all of that and the happier everybody is. That data is helping to inform our roadmap for this coming quarter. We're spending a lot less time on tooling, a lot more time on direct impact. I feel a lot more rewarded personally. If I had to go do it all over again and rewind the clock a few years, I would have made a harder push to onboard tools like this sooner than we did,” says Ryan.

Visit Fivetran’s Financial Analytics page to learn more about automating your financial reporting.

Visit Hightouch’s website to learn more about activating your data warehouse and putting data in the hands of customer facing teams.

Kostenlos starten

Schließen auch Sie sich den Tausenden von Unternehmen an, die ihre Daten mithilfe von Fivetran zentralisieren und transformieren.

Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.
Company size
Region
Industry
Financial services & insurance
Sources
No items found.
Destinations
No items found.
Sources
No items found.
Destinations
No items found.
BI tool
No items found.
Try Fivetran for free
Key results
The total economic impact of Fivetran

Learn how automated data movement boosts productivity and accelerates insights for your business.

Download the report
Are you a true data pioneer?

Our quiz will tell you where you are on your data journey and what tools you can unlock to level up.

Take the quiz
Why they chose Fivetran

Further reading
No items found.
No items found.
Related customer stories
Case study

National Australia Bank enhances customer experiences and powers GenAI

Case study

Envision Pharma Group and Fivetran: Pioneering faster paths in pharma

Case study

Raiffeisen Bank International uses real-time data to win more customers

Case study

Saks achieves data efficiency & enables AI with Fivetran

Case study

Cemex connects 1,800+ global facilities in real-time

Case study

Gill Capital unlocks SAP data with Fivetran, boosts sales by 25%

Case study

HubSpot spart bei GenAI mit Fivetran 100.000 US-Dollar ein

Case study

Deliveroo verwandelt die Essenslieferung in eine datengesteuerte Unternehmung

Case study

Paylocity transforms its go-to-market strategy with Fivetran

Case study

Group 1001 increases productivity by 10X

Let's get you moving data