The Looker platform is the ultimate solution for self-service analytics and business intelligence. The platform combines a version-controlled modeling layer called LookML with a self-service query feature that business users love. And best of all, the tool is designed to minimize the trivial and repetitive tasks that slow down data analysis. However, although Looker introduced a paradigm shift in the BI space, it can be a double-edged sword if managed improperly.
If Looker is implemented haphazardly, you’ll end up two steps back with code and model clutter and metrics that won’t match across reports.
When implemented correctly, your Looker instance will provide a trustworthy source of truth, offering self-support from business users.
I’ve lived in the analytics and business intelligence space for nearly two decades spanning enterprise technology and startups. I was also head of internal analytics at Looker and saw them through their exit to Google. When I come in to help Looker customers, it’s usually for one of several reasons.
Here are the five most important signs it’s time to hire a Looker consulting partner.
You should hire a Looker consultant when…
Your LookML is a mess
The most frequent issue I see is that an instance doesn’t separate business logic between the LookML and content thoughtfully. A few signs of this are overuse of derived tables, redundant explores, and unnecessary LookML that could have been avoided with better leverage of the content layer. This leads to a lack of metric integrity, excessive analytics oversight, and ultimately a lack of trust in Looker reporting company-wide.
Your Looker instance lacks architecture
Messy LookML is more often a symptom than a cause. Likely, analysts aren’t following an architectural plan, or worse; there wasn’t one to begin with.
Good architecture starts with empathy. Having empathy for the business users helps identify the type of questions that will be asked. This ultimately leads to a better self-serve environment. The goal is to create a minimal number of explores that answer the broadest range of data requests in an experience tailored to the data consumer.
Analysts fear breaking the LookML model
Nothing slows down progress like the fear of breaking something. This is especially common for Looker developers who support embedded analytics. A poorly maintained Looker model will deter developers from embellishing existing logic or making needed corrections. This results in “coding around the problem,” which exacerbates current issues instead of addressing them.
It’s your first Looker implementation
When the LookML model is fresh, it’s easy and exciting to spin up brand-new content for your business. But Looker isn’t a simple data visualization tool. Without foresight, you’ll find yourself having to rebuild your model. This breaks newly created content and can undermine trust right from the start.
Implement best practices early! Especially when you are migrating to Looker from another BI tool, you’ll want a customized migration plan. Thoughtful planning will ensure that your dashboards and reports are consistent with prior versions. Your success will be measured by how many business users continue to engage with the content.
Your analysts need real LookML training
If you lack a good LookML developer, your Looker model can stagnate and inhibit adoption. When this occurs, a good Looker consultant can “right the ship” by training or helping hire new Looker developers. Co-development sessions are the best avenue to get new or existing Looker analysts up to speed. I encourage analysts to have agency over the Looker model and absorb best practices by working alongside a consultant.
If you skipped the intro, you’ve probably guessed that I am a Looker consultant. I provide remedies to all of the issues I’ve listed above — from development to analytics to training and support. I love helping organizations get the most out of their data by providing self-serve analytics for every business unit. So if any of these points resonate with you, I’d love to chat.
What does a Looker consultant do?
As a Looker consultant, I can tell you that no two engagements are identical. However, I usually provide a mix of a few things:
- Data strategy — a holistic perspective of how a business can apply data and technology to its strategic and tactical decision making
- Software expertise — a Looker certified professional with an understanding of the pitfalls to avoid and how to avoid them
- Data architecture and data modeling — expert data warehouse, ETL, and LookML guidance to demonstrate the best way to deliver data products and the data that is required to produce them
- Data expertise — understanding of the intricacies of common but complicated data sources like Netsuite, Salesforce, or Google Analytics
- BI expertise — the ability to translate business questions into analyses that lead to maximum insight with minimal overhead
- Data governance: ensuring that data is properly governed by establishing access controls, implementing data privacy measures, and adhering to regulatory requirements
- Looker training — education spanning from end-user training to help business users master the Looker UI to developer training and co-development
What does a consulting engagement look like?
Every client I’ve talked to has the same questions regarding hiring consulting services. Essentially, they all boil down to what will I get and how much will it cost? I maintain a transparent approach to my consulting practice, so allow me to share what an engagement typically looks like.
Engagement Scope
The scope of every engagement shares the same general objectives. 1) I aim to provide data to the end user in a way that they can most effectively consume it. 2) Implement a solution with the cleanest architecture and most efficient code possible. Reaching this objective depends on a few things.
How “deep” in the data stack does the project need to start? Sometimes the data is prepared for data modeling. Then the project is just a matter of superficial metric design and business analytics. Other times, the data is all over the place, or it may not be stored at all. Those cases will require a consultant data engineer to build a data platform just to get the data to a place where it’s ready for modeling.
How many internal resources are available on the project? An engagement can range from the augmentation of an existing data team to fractional data leadership and outsourced execution. You can imagine how each of these scenarios affects a project’s scope.
Is there an established perspective on data and analytics? Beyond the technical and resource requirements, there is business consulting to help merge the impact of data with business decisions and processes. Helping businesses understand their data strategy is a whole different proposition than providing a few financial metrics here and there. Each demand has its value, but the inputs are significantly different.
The factors listed above impact the varying size of an engagement, but the shape of each consulting engagement is usually similar. Engagements usually start with a well-defined project to solve a problem or achieve a goal. From there, the level of consulting involvement tapers over time as clients upskill and take ownership of their Looker instance. With the tapering involvement comes tapering rates.
Consulting Rates
As I mentioned above, the size of engagement will impact the overall rates. As is common, more hours per week over more weeks will reduce the cost of consulting hours. Conversely, fewer hours over fewer weeks will increase the price hourly rate. Retainer rates that extend beyond initial projects usually benefit from the size of the scope of the initial engagement.
Most of all, consulting rates vary based on experience. An experienced Looker consultant will charge anywhere from $150 to $250 per hour, depending on the scope of an engagement.
It’s possible to find freelance Looker developers picking up side jobs on UpWork for less than that. But hire these freelancers at your own risk. They’re often prone to creating one of the problems listed above.
On the other hand, larger consultancies will charge higher fees because of their agency overhead. Be careful not to pay expert consulting rates for work passed to a junior analyst within a consultancy.
Finally, work that is either highly strategic or highly technical (data engineering and data modeling) will generally fetch a premium. But again, the fundamental value of this work is worth paying for to avoid the situations listed above.
A final note from me
I hope this clarifies the whole consulting proposition. The upside of bringing in a certified expert can be pivotal. It can mean the difference between reaching your business’s potential, backed by data, and a frustrating cycle of confusion about untrusted data.
If you want to learn more about Looker and the rapidly transforming data space, please click the “follow” button. And don’t hesitate to get in touch on LinkedIn or contact me for consulting services.