Screenshot 2017-09-06 10.30.08.PNG

Our take on Looker: “Modern analytic technology for data product teams”

Overview

When you look at the analytic landscape, it becomes apparent that most platforms are either directly focused on enterprise analytics or are derived from platforms intended for use inside a businesses walls. 

It’s rare to find a platform that works both for internal analytics and for data product teams. Looker is that rare platform that, while initially focused on internal use cases is an excellent choice for those seeking to build data products with embedded analytics.  

Looker’s cloud-based approach offering fast, flexible, and powerful analytics allow data product teams to focus on creating value for their customers rather than on the care and feeding of analytical systems.

Who’s It For?

Product teams tend to fall into one of several categories: 

  1. Teams that want a toolkit to build their own customized analytics and to save cost and are willing to maintain the platform entirely in-house,

  2. Teams that want flexibility but will accept limited customization capabilities in order to have the platform managed for them, and

  3. Teams that want everything done for them and don’t mind the costs associated with a fully-managed analytic platform. 

Looker is an ideal platform for those in category 2—they want flexibility but still want the benefits of a full-featured analytic platform.

Those looking for a “data product toolkits” that allows every aspect of the user interface to be completely customized may be better served by one of the numerous build-it-yourself offerings on the market, but I think this is a limited set of data product buyers. Looker offers deep customization capabilities that should satisfy all but the most obsessive data product builders.

Similarly, if you need a product to be full-managed—completely hands-off for your team—Looker might not be the best choice for you. The business model assumes that you have at least some technical resources available for development and continued maintenance of the platform.

With this said, I feel that Looker is one of the best choices available for most teams building a data product. It’s fast, flexible, and customizable and the user interface is a pleasure to use both for builders and for end users. 

They’ve considered the on-going product management needs to the degree that is seldom found in other platforms and the team truly understands data products.

Key Features

Looker is completely web-based and offered in a software-as-a-service (SaaS) model. You aren’t required to download Looker and install it on your own machines (though that’s an option); the platform is managed by Looker and connected to your product via APIs or embedding as iFrames.  

This is a great mode of operation for the majority of data products, allowing product teams to focus on product design and management instead of infrastructure support.  

However, if you insist on having Looker installed on-premise, they can accommodate that model as well. Same price, same functionality, just managed by you instead of by the Looker team.

With Looker, you can create embedded analytics for your data product in multiple ways.

  1. Embedded as an iFrame. When you add Looker to your data product using this method, you can offer the entire Looker experience—the entire user interface—to your users within your core application. It might be a tabbed section of your product that allows the user to use a fully white-labeled instance of Looker within your application. This can have functionality restricted or expanded as needed. Want your user to use Looker as a playground to explore data at will? You can do that. Want to show charts and tables in a specific format that users are unable to modify? You can do that too with Looker embedded as an iFrame within your product.

  2. Embedded alongside workflow. You can also embed elements of Looker—a chart, a table, an infographic or just a metric value—right alongside your product’s workflow.

    Say you’ve got a system that allows customers to manage inventory for multiple locations. You can embed inventory level charts, quantities remaining, etc., right next to your functionality for placing new orders. This ability to offer Looker-derived elements in context is exceptionally powerful. And, since Looker manages the definitions of all of your metrics, you don’t need to worry that “amount remaining” is different from page to page, user to user.

    The embedding experience is fantastic. It’s so good that the Looker demo is set up so that it shows you which elements of the data product were built by the “customer” and which are powered by Looker. The experience is so seamless that, if they didn’t point out which was which, you’d never know where one ended and the other began.

If you watch a demo of Looker (and you should—it’s impressive) you might hear the phrase “modern analytics for modern data products.”  What’s this mean?

For those building data products, it means an entirely different architecture from older platforms that move data into data marts before the analytics can be displayed. Looker takes advantage of the speed and functionality of modern databases (such as Snowflake or Redshift) and acts on your data in place, rather than copying a subset to a storage location.

This is huge for data products. 

First, it means you don’t need to worry about moving, synchronizing, and paying for storing data a second time.

Next, it means that you can drill down from high-level aggregate charts (like quarterly averages) all the way down to the source, transaction-level data. With platforms that store extracts of your data, this is tough to do and is usually only done when the cost and extra work are warranted. With Looker, the full drill down capability is built-in and incurs no extra costs or time.

Finally, the use of modern database functionality allows Looker to operate in real-time if needed. While systems that extract and store data for analytics often refresh once or twice per day, Looker is operating on your source data. Need to know what’s happening this very second?  Looker is the platform for you.

But in addition to these capabilities that result from the use of modern database technologies, Looker has a few more tricks up its sleeve…

With many platforms, each customer of your data product will have its own data store. Add a new customer?  Add a new data store. With Looker it’s different. Since data isn’t moved into a separate data store for each customer, the on-boarding of a new customer is simple and done through Looker’s robust API. Think hours versus weeks.

This also brings tremendous benefits when you are structuring your data product offering. Many teams choose to offer “Basic,” “Plus,” and “Pro” versions of their analytics but this can be difficult on some platforms. It can mean changing data models and storage schemes—not always simple tasks. With Looker, it’s a matter of sending a parameter to a customer that identifies theme as basic, plus or pro. Send the Pro parameter, and they receive the right dashboard template, the right permissions, and the right functionality for the desired access level. It makes the life of a product owner that much easier. The value of this method of controlling the product fearer set became apparent when I was able to see a recent demo of Looker’s embedded analytics capability. During the session, they switched the mode of the data product from “Basic” to “Premium” with the click of a button. One click and BOOM—new features appear on the analytic dashboard. Impressive and much more difficult to do with most platforms. If you’re considering using a tiered data product offering, you should pay close attention to this benefit.

Services

Services—both initial deployment and on-going support—can make or break a data product. Often product teams are expert in their core application, but lack experience building a data product with embedded analytics. This is where professional services come into play and Looker excels in this area. Every project includes a “jump start” component where Looker’s services team helps you define your data product, from data model to dashboard design. As someone who has seen even the most talented teams struggle for a year or more trying to get an initial data product launched, I can tell you—a jump start program such as the one Looker offers in essential to success. The jump start program is run in-house by Looker’s services team rather than out- sourced to partners.

Support doesn’t end with initial implementation. Looker prides itself on ongoing customer support. During my discussions with Looker, they revealed an amazing statistic: the average response time for a customer question via chat is 24 seconds. Twenty. Four. Seconds. Amazing. And, this isn’t from an untrained, inexperienced call answerer. It’s from their professional support team who can answer your question right there in the first chat. None of this “I’ll have to pass that on to the Engineers stuff.”

If you’ve built mission-critical customer analytics and something isn’t going right, you want that kind of instant support.

Pricing

Pricing a platform for building data products can be difficult. Too high a price and you make it tough for the product team to offer analytics at a reasonable price point. Too low and the initial investment in your platform is never recovered.

Looker’s standard pricing model is “platform plus users.” This means that you pay a fee for the use of the platform and then an additional amount for each person using the Looker platform for the data product. If your data product will have 10,000 users, you can expect to pay the platform fee + (the per user fee)*10,000 to use Looker to power your product.

To be frank, this poses a problem for data product leaders. Often, it’s unclear how many users you will attract early in the process. And, if you want to attract as many users as possible to the data product—perhaps by starting with a free tier and encouraging customers to buy more paid premium analytics—a per user fee is a problem. It naturally reduces the incentive to bring users—especially at the lower introductory price tiers—to the product. 

Looker understands this. When I asked the Looker team about pricing models, they were very clear that they understand the widely varying needs of data product teams and are willing to create a pricing model that makes sense for your situation so that you can offer analytics profitably. If I were deciding which analytics platform should power my data product, I wouldn’t let the “platform plus users” pricing model dissuade me from using Looker. They get it. They’ll work with you.

Caveats

There must be a few gotcha’s, right? Is Looker perfect for everyone? Of course not. Although Looker is an excellent platform on which to build a data product, there are some caveats.

If you have absolutely no resources available for your data product, to create a strategy for your analytics, to design metrics, to help find and connect data sources, to create analytics that will engage your users, to support the product once launched, you might want to consider a different platform. Looker isn’t a “sign the contract and walk away” solution for your data product. They require a degree of participation on your part. They assume that you have at least some resources available, both for initial strategy and implementation and for on-going support. They jump start the process and will provide support throughout the lifecycle, but they do need guidance and participation from your team. 

If you need to extensively customize the user interface (beyond background colors, chart colors, etc.), you might want to choose another option. Looker has a great set of customization capabilities (available through the user interface) but lacks the ability to tweak every little setting. If you insist on 5 point corner radiuses—not 2, not 8—on select charts, you might be better off elsewhere.

Looker relies on the analytic power provided by modern database such as Snowflake, Redshift, or BigQuery. If you are using an older SQL database technology, Looker will still work, but you might not get the full benefit of it’s capabilities. In this case, I’d recommend upgrading the database and then using Looker to get the maximum impact.

Finally, if Looker’s embedding API and iFrame model isn’t enough and you require technologies like javascript for embedding, Looker isn’t the best choice for you.

Final Thoughts

Building a successful data product using embedded analytics is very different from building inside-the-organization analytics. 

It requires functionality that's different, strategy that’s different, and pricing models that are different. Looker understands these needs. 

It’s a modern analytic platform that takes advantage of the power of today’s databases and allows product teams to create capabilities that were difficult just a few years ago. Drilling all the way down to individual records, view data that is up-to-the-second fresh, and the rapid deployment to new customers are all capabilities that will benefit buyers choosing Looker.

A completely web-based authoring and management environment make life easy for product teams and eliminate the need to distribute authoring tools to power-user customers as is required by some platforms. And all of these capabilities are back by an incredible team that’s experienced and ready to help your get your data product running (and stay running). 

Looker is an excellent choice for teams seeking to build powerful, flexible, and user-friendly data applications.

 

Want to download this information as a PDF?
Get it free from our premium content center.