Keeping Your Customers Informed
How we used Looker to transform our customer reporting
I joined Broadlume two years ago, back when it was a company called AdHawk in the AdTech sector. Todd Saunders, our CEO, approached me with a simple problem: he needed help using all the data we had to tell the story of the company’s performance.
At the time, there was certainly a lot going on at AdHawk. We were transitioning from the AdTech business that the company was founded on, to a technology leader in the flooring space. Helping Todd solve his data problem not only helped our business, but helped us create a solution for all the flooring businesses we support.
The problem Todd had wasn’t unique. From large corporations to small businesses, the people managing the business often struggle to marry all the data they are collecting into a cohesive picture. It’s also an ever evolving problem. As soon as you have a handle on your business, you may introduce a new product or service that will change the underlying data set and make the reporting useless. The field of Business Intelligence (BI) is a fast growing one due to this problem and its ever evolving nature.
To help solve Broadlume’s data problem, I used a tool set I knew. I had spent the previous five years at a startup called Stack Overflow where I had used Looker as the primary reporting tool. Looker was the first BI tool I used, and my experience learning in the space grew as the tool became more powerful with feature enhancements. I even got to participate in a power user marketing campaign with Looker, resulting in this silly billboard.
The Foundation: Our BI Stack
Having the experience with Looker, I got to building out the backbone of Broadlume’s reporting. I started by using Fivetran, a data pipelining tool, to collect information from our various sources and send it to our main data warehouse, Redshift. Our data stack looked something like this:
With the foundation set up, I set forward on creating the proper connections between the various data sources and setting up reporting. This reporting started to help us better understand our business performance. From web analytics to customer growth and churn, we now had the right data to help us make decisions.
Problems With Our Existing Reporting Tools
At the same time I was creating our data infrastructure, we were running into a problem with the reporting we were providing our customers. Our main product is a website platform that offers our flooring customers a way to showcase their products. We also offer digital advertising services to help drive traffic to that website. For our customers, understanding the analytics performance of the website and leads generated by the digital advertising spend is very important.
To help provide insights to our customers, we were using a software that helped us connect some data sources (Google Analytics, Google and Facebook Advertising, Call Tracking Metrics, etc.) together to paint a portrait of performance. This software would send a pdf report each month to our customers that showcased their performance. This was fine enough when we had one website platform and a couple advertising products, but as we started to grow and expand rapidly we ran into a couple issues:
- Frequency — this report would only run once a month. Due to that, there was a lag in understanding the outcome of change, leading to performance issues we would only see after the fact.
- Flexibility — there was only one type of report. If you wanted to add a new product, you got a new report. Some customers would get several reports a month that weren’t connected, leading to confusion around performance.
- Breakage — if something about the report didn’t work, it would still send, but show all 0s for performance. No Account Manager likes getting an angry email from a customer referencing a report showing 0 in performance across the board.
The Solution: Looker
While discussing these problems and a potential solution, a simple question kept popping up. “Why can’t we provide our customers with the same reporting we use internally?” The answer was simple enough, we could! Looker offers a tool called Embedded Analytics which allows you to embed Looker dashboards into your own platform. This solution seemed ideal for us and fixed the above problems:
- Frequency — Looker reporting will be in real time
- Flexibility — We were in control of the data and dashboards, we could constantly iterate and test
- Breakage — We could see breakage in real time and fix any issues ASAP
DealerHQ: Our Reporting Home
Lucky for us, we had a platform ready to use. As part of being a customer of Broadlume, you get access to our DealerHQ platform. Two years ago, this platform was in its early stages, the main functionality was tracking leads that came in to the customer’s website. This felt like the ideal location to include reporting on website analytics, leads, and advertising performance.
We got to work creating the basis of our reporting. We quickly discovered two key advantages to using Looker for this purpose. First, we had the data already set up internally, so we just had to mimic that setup for our customers. Second, setting the product up wasn’t a strain on resources. It required a couple weeks of engineering time to set up the initial build, after that it was building with Looker as usual.
Within a couple of months, we had three dashboards ready for our customers to use. Here’s an example of what a dashboard looks like in our DHQ platform:
Over a year after launching reporting in DealerHQ, we’ve been able to take a step back and look at what we’ve learned. Here are some insights:
- Iteration is key. We often start with one variation of a report and change various small aspects before it gets to our user base. After that, we solicit feedback from users and continue to change the product. Looker makes it easy to test and change features quickly.
- You can do a lot with a little. Currently, we have a portion of a couple teams’ time to build and maintain this reporting. Using Looker helped us produce the reporting we needed without dedicating full teams to the project.
- Keep it simple. Our users are running a business. The reporting we give them shouldn’t bog them down. It should be straightforward and to the point, allowing them to understand the performance of the products they are paying for. Some customers will be power users and others will barely look at the tool, but it’s important to keep the common user in mind when creating the dashboard.
With over a year of reporting under our belts, we’ve got some big ideas for the future. We just hired our first Business Intelligence Analyst to help take us to the next level. We also have some exciting new products to build reporting for in the pipeline!
