Friday, July 1, 2022
HomeBig DataWay forward for the Metrics Layer with Drew Banin (dbt) and Nick Handel...

Way forward for the Metrics Layer with Drew Banin (dbt) and Nick Handel (Rework) – Atlan


Scorching takes on what we get fallacious in regards to the metrics layer and the place it suits within the fashionable knowledge stack

The metrics layer has been all the fashion in 2022. It’s simply forming within the knowledge stack, however I’m so excited to see it coming alive. Not too long ago dbt Labs integrated a metrics layer into their product, and Rework open-sourced MetricFlow (their metric creation framework).

A number of weeks in the past, I used to be fortunate sufficient to speak in regards to the metrics layer with two most prolific product thinkers within the house — Drew Banin (Co-founder of dbt Labs) and Nick Handel (Co-founder of Rework).

We coated every part from the fundamentals of a metrics layer and what folks get fallacious about it to real-life use instances and its place within the fashionable knowledge stack.

Earlier than we start… WTF really is a metrics layer? At the moment metrics are sometimes cut up throughout completely different knowledge instruments, and completely different groups or dashboards find yourself utilizing completely different definitions for a similar metric. The metrics layer goals to repair this by creating a typical set of metrics and their definitions.

Drew and Nick dove extra into this definition, so let’s bounce proper into all of their insights and fiery takes. We talked for over an hour, so it is a condensed, edited model of our dialogue. (Take a look at the total recording right here.)

How would you clarify the metrics layer to a newbie knowledge analyst?

Because it’s a brand new idea, there’s a variety of confusion about what actually the metrics layer is. Drew and Nick reduce by means of the confusion with succinct definitions about creating a typical supply of reality for metrics.

Drew Banin: “The shortest model I can consider is…”

Outline your metrics as soon as and reference them in all places in order that in case your metrics ever change, you get up to date outcomes in all places you have a look at knowledge.

Nick Handel: “The best way that I’ve defined it to household and people who find themselves completely out of the house is simply, companies have knowledge. They use that knowledge to measure their operations. The purpose of this software program is mainly to make it very easy for the info analysts (the people who find themselves accountable for measuring that knowledge) to outline these metrics, and make it straightforward for the remainder of the enterprise to eat that single right strategy to measure that knowledge.”

What’s the actual drawback the metrics layer is seeking to clear up?

Nick and Drew defined that the metrics layer is motivated by two key concepts: precision and belief.

Nick: “I believe we’re all fairly satisfied in regards to the worth of knowledge. We’ve every kind of various, fascinating issues that we will do with knowledge, and the price of doing these issues is pretty excessive. There’s a bunch of labor to get the info into the place the place we will go and do something that’s actually fascinating and helpful.

“Why does this matter? It’s purported to make that entire means of getting the info prepared for that supply of worth a lot simpler and likewise extra reliable.”

It comes all the way down to these two issues: productiveness and belief. Is it straightforward to supply the metric, and is it the fitting metric? And might you set it into no matter software you’re attempting to serve?

Drew: “That’s actually good framing. I simply look inwards at our group. The very first metric we ever created was weekly lively initiatives — what number of dbt initiatives had been run within the earlier seven days? Now we’re about 250 folks and we’re measuring so many issues throughout the enterprise with a lot of new folks round.”

We’re attempting to make it possible for when somebody says ‘weekly lively accounts’ or ‘MRR’ or ‘MRR cut up by handle versus self-service’, all of us imply precisely the identical factor.

Drew and Nick additionally emphasised change administration as each a significant problem and use case for the metrics layer.

Drew: “I believe a lot in regards to the change administration a part of it. If you happen to get the fitting folks collectively, you possibly can exactly outline a metric at that cut-off date. However inevitably your corporation or product will evolve. How do you retain it in sync in perpetuity? That’s the laborious half.”

Nick: “I actually agree with that. Particularly if change administration is going on when there are just a few folks within the room, and different people who find themselves relying on the identical metrics weren’t part of that dialog.”

How ought to we take into consideration the metrics layer, and the way ought to it interaction with different elements of the trendy knowledge stack?

Nick broke the metrics layer down into 4 key elements (semantics, efficiency, querying, and governance), whereas Drew targeted on its function as a community connecting a various set of knowledge instruments.

Nick: “The best way that I take into consideration the metrics layer is mainly 4 items. There are the semantics: How do I’m going and outline this metric? This will vary from ‘Right here’s a SQL snippet’ or ‘That is the definition of the metric’ to a full semantic layer that has entities and measures and dimensions and relations.

“Then there’s efficiency. Nice, now I’ve this semantic mannequin. How do I’m going and construct logic towards it, executed towards some compute setting (whether or not it’s a warehouse or only a compute engine on an information lake)?

“Then there’s, how do I question this factor? What are the interfaces that I take advantage of to drag it out of the info warehouse or knowledge lake, resolve it into this quantitative object that I can then go and use in some evaluation. That features each broad methods of consuming knowledge (like a Python interface or GraphQL or a SQL interface) in addition to direct integrations (a device that builds a customized wrapper round a REST or GraphQL API and builds a very firstclass expertise).

“Then the final piece is governance. There’s organizational governance and technical governance. Organizational governance that means, does the finance chief agree on the human-understandable definition of income in the identical approach that the technical one that’s defining the logic defines that code?”

Drew: “Simply to supply an alternate framing: We will consider it when it comes to the expertise for the one who needs to eat knowledge to reply some query or clear up some drawback, after which additionally the folks constructing the instruments the place these of us are consuming the info.

“It’s a little bit bit at odds with one another, as a result of the enterprise customers need to see the very same metric in each single device they usually need all of it to replace in actual time. So you’ve this large community of various instruments that conceivably want to speak to one another. That’s a tough factor to prepare and make occur in follow.

That’s why the concept we name this the ‘metrics layer’ is sensible. It’s a single abstraction layer that every part can interface with in an effort to get exact and constant definitions in each single device.

“To me, that’s the place metadata actually shines. Like, that is the metric, that is the way it’s outlined, that is its provenance, right here’s the place it’s used. This isn’t really the info itself. It’s attributes of the info. That’s the knowledge that may synchronize all these completely different instruments collectively round shared knowledge definitions.”

What metadata ought to we be monitoring about our metrics, and why?

Nick and Drew shared that metadata is essential for understanding metrics as a result of firms lose essential tribal data about knowledge outages and anomalies over time as employees modifications.

Nick: “The metric is among the most constant objects in a company’s life.

Merchandise change, tables change, every part modifications. Even the definitions of those metrics evolve. However most companies find yourself monitoring the identical North Star metrics from the very early days. If you happen to can connect metadata to it, that’s extremely beneficial.

“At Airbnb, we tracked nights booked. It was essential from the very early days when BI was actually a printed-off graph that they placed on the wall, and it’s nonetheless an important metric that the corporate talks about within the public earnings calls. If we had been monitoring essential metadata by means of time of what was occurring to that metric, there could be a wealth of information that the corporate may use.”

They defined that these modifications are why it’s essential for the metrics layer to work together with each the info layer and the enterprise layer — to seize context that impacts knowledge evaluation and high quality.

Nick: “Airbnb had a giant product launch, and completely different metrics spiked in all completely different instructions. At the moment, I’m unsure {that a} knowledge scientist at Airbnb may actually perceive what occurred. They’re attempting to make use of historic knowledge to know issues, they usually simply don’t have that context. If something, they actually solely have context for the final two or three years, when there was any individual who’s within the enterprise who remembers what occurred, who did the evaluation, and many others.”

Drew: “There’s a variety of this that finally ends up being technical — when it comes to how instruments combine with one another, and the way you outline the metrics and model them. However a lot of it’s certainly the social and enterprise context.

In follow, the folks which were round for the longest time have essentially the most context and doubtless know greater than any of our precise methods do.

“We had a interval the place we had a little bit bit of knowledge loss for some occasions we had been monitoring. It appeared like, I believe it was, Could 2021 was the worst month ever. However actually it was identical to, no, we didn’t accumulate the info.

“How would you already know that? The place does that data dwell? Is it a property of the supply dataset that propagates by means of to the metrics? Who’s accountable for encoding that?”

What are the actual use instances for a metrics layer?

Drew and Nick referred to as out a variety of potential functions for the metrics layer — e.g. bettering BI and analytics for early-stage knowledge groups, serving to enterprise and knowledge folks use knowledge fashions in the identical approach, and making beneficial however time-consuming functions (like experimentation, forecasting, and anomaly detection) doable for all firms.

Drew: “I believe among the use instances round BI and analytics are essentially the most clear, apparent, and current for lots of firms.

Many firms on the market aren’t on the knowledge science and machine studying a part of their journeys but. Issues that make enterprise intelligence and reporting higher (extra exact and extra constant) cowl 90% of the issues that they’re attempting to resolve with knowledge.

“Casting our minds ahead, I believe that there may very well be a ton of advantages to leveraging metrics for knowledge science use instances.

“Particularly, one of many issues that we’ve seen folks do with dbt that was actually formative for me — they might construct these knowledge fashions after which use them each for BI reporting and likewise to energy knowledge science functions and modeling. The truth that the info scientist and the BI analysts are utilizing the identical knowledge units signifies that it’s much more probably that they’re consuming the identical knowledge in the identical approach. If you prolong it to metrics, there’s like a very pure strategy to make that occur too.”

Nick: “I do partly agree with that. But additionally there are a variety of knowledge science and machine studying functions that require very completely different datasets than what a metric retailer produces.

“In analytics functions, you attempt to embody as a lot related data as doable. If in case you have an ecommerce retailer, folks can browse it logged out. So that you attempt to dedupe customers and establish as customers log into units. There’s a complete follow of attempting to determine which entities are utilizing your service. That’s actually essential for analytics as a result of it permits us to get a a lot clearer image. However you don’t need to try this for machine studying, as a result of that’s all data leakage and that can destroy your fashions.

With machine studying, you attempt to get as near the uncooked knowledge units as doable. With analytical functions, you attempt to course of that data into the clearest and finest image of the world.

“One of many functions that I all the time take into consideration is experimentation. The rationale we constructed a metrics repo initially was experimentation.

“There have been 15–20 folks on the info staff on the time. We had been attempting to run extra product experiments, and we had been doing every part manually. It was actually time intensive to go and take project logs and metric definitions and be a part of them collectively.

Principally, we would have liked some programmatic strategy to go and assemble metrics. It’s a vastly beneficial software for firms that do it, however only a few firms have the infrastructure or construct the tooling to do that. I believe that that’s actually unlucky. And it’s in all probability the factor that I’m most excited in regards to the metrics layer.

“If you consider each knowledge software as having some value and a few profit — the extra you possibly can scale back the price of pursuing that software, the extra clearly the justification turns into to pursue some new software.

“I believe experimentation is one among these examples. I additionally take into consideration anomaly detection or forecasting. These are issues that I believe most firms don’t do — not as a result of they’re not beneficial, however simply because producing the datasets to even get began on these functions is admittedly laborious.”

Let’s bounce into some questions in regards to the metric layer and the trendy knowledge stack.

First, let’s discuss bundling vs unbundling. Ought to the metrics layer even be a separate layer, or ought to or not it’s a part of an present layer within the stack?

As with each debate within the knowledge ecosystem, we ended up simply answering, it relies upon. Drew and Nick defined that how we clear up this drawback is finally extra essential than how we outline that answer.

Drew: “I’m not in love with the way in which that we as an ecosystem discuss new instruments as being layers, just like the lacking layer of the info stack. That’s the fallacious framing.

“Those that construct functions don’t give it some thought that approach. They’ve companies, and the companies can discuss to one another. Some are inside companies and a few are SaaS companies. It turns into a community of related instruments slightly than precisely, say, 4 layers. Nobody runs an software anymore with precisely the Linux, Apache, MySQL, and PHP (LAMP) stack, proper? We’re previous that.

The phrase ‘layer’ is sensible solely insofar because it’s a layer of abstraction. However in any other case, I reject the terminology, though I can’t consider something too significantly better than that.

“The very last thing I’m going to say on bundling and unbundling… For this factor to work, it does must be an middleman between a really large community of various instruments. Treating it as a boundary like that motivates which instruments can construct it and supply it. It’s not one thing you’ll see from a BI device, as a result of it’s not likely in a BI device’s curiosity to supply the layer to each different BI device — which is just like the factor that you really want from this.”

Nick: “I believe I typically agree with that.

Principally, folks have issues, and firms construct applied sciences to resolve issues. If folks have issues and there’s a beneficial expertise to construct, then I believe it’s price taking a shot and attempting to construct that expertise and voicing these opinions.

“In the end, I believe that there are good factors there of the connection to completely different organizational workflows. This isn’t one thing that I believe we’ve finished a very good job of explaining, however I believe that the metrics retailer and the metrics layer are two completely different ideas.

“The metrics retailer extends the metrics layer to incorporate this piece of organizational governance — how do you get a bunch of various enterprise customers concerned on this dialog, and truly give them a task in one thing that, frankly, they’ve an enormous stake in? I believe that that’s one thing that’s not actually caught on this dialog across the metrics layer, or headless BI, or any of those completely different phrases. However it’s actually, actually essential.”

For a conventional firm that already has an information warehouse and BI layer, the place does the metrics layer match into their stack?

Once more, the reply is that it relies upon — sigh. The metrics layer would dwell between the info warehouse and BI device. Nonetheless, each BI device is completely different and a few are friendlier to this integration than others.

Nick: “The metrics layer sits on prime of the info warehouse and mainly wraps it with semantic data. It then permits completely different endpoints to be consumed from and mainly pushes metrics to these completely different locations, whether or not they’re generic or direct integrations to these instruments.”

Drew: “It finally ends up being very BI device–dependent. There are some BI instruments the place it is a very pure sort of factor to do, and others the place it’s really fairly unnatural.”

If an organization has already outlined a ton of metrics inside their BI device, what ought to they do?

Nick and Drew defined that gradual and regular wins the race if you aren’t ranging from scratch. As an alternative of planning an enormous overhaul, begin with one staff or device, combine a greater metrics layer, and take a look at the way it works in your group.

Nick: “I’d advocate for not an enormous ‘change every part ’. I’d advocate for, outline some metrics, push these by means of the APIs and integrations, construct one thing new, probably exchange one thing outdated that was laborious to handle, after which go from there when you’ve seen the way it works and consider in that philosophy.”

Drew: “I’m with you. I believe one thing domain-driven makes a variety of sense. You may validate it after which increase. I’d in all probability begin with… it is determined by your tolerance, however the government dashboard that goes to the CEO. Is that the perfect place to kick the tires? Perhaps not. But when it really works there, it’ll work in all places.”

Can’t a metrics layer simply be a part of a function retailer?

Since Nick has constructed a number of function shops and metrics layers, he had a robust opinion on this matter — whereas the metrics layer and options retailer are related, they’re too essentially completely different to merge proper now.

Nick: “I’ve a very robust opinion about this one as a result of I’ve constructed two function shops and three metrics layers. These two issues are completely completely different.

“On the core, they’re each derived knowledge. However there are such a lot of nuances to constructing function shops and so many nuances to constructing metric shops. I’m not saying that these two issues won’t ever merge — the thought of a derived knowledge repository or one thing like that sounds great. However I simply don’t see it occurring within the quick time period.

Everybody needs options to be particular to their mannequin. No person needs metrics to be particular to their staff or their consumption. Individuals need metrics to be constant. Individuals need options to be distinctive and no matter advantages their mannequin.

“Actual-time versus batch — it is a tremendous difficult drawback within the function house. Organizational governance is approach essential for the metrics layer. The technical definitions are sometimes completely different. The extent of granularity is completely different for options — you go approach finer with options than you do metrics.”

Do you consider a caching layer is crucial for a metrics layer?

This was a powerful YES from each Drew and Nick. Caching makes the metrics layer quick, which is crucial for guaranteeing that knowledge practitioners really use it. Nonetheless, it’s essential that this caching doesn’t replicate knowledge.

Drew: “I believe that the velocity with which you’ll be able to ask a query and get a solution again is admittedly crucial.

The distinction between one thing taking a minute plus to come back again and never coming again in any respect is negligible in a variety of instances. So, conceptually, I’m very aligned with the thought of caching metric knowledge and having the ability to serve it up actually rapidly.

“I’ll simply say — and I believe we’ve been open about this previously — we in all probability gained’t try this for V1 of metrics inside dbt. However conceptually, I’m fairly aligned with that being an essential a part of the system long-term.”

Nick: “Caching is tremendous essential. Efficiency issues a ton, particularly to enterprise customers. Even 10 seconds is lower than a really perfect expertise.

“I believe that there are two essential nuances to caching. One is, what do I do know forward of time that I would like, and the way do I pre-compute that and make that actually snappy? After which if I do compute one thing, how do I then reuse it in order that it’s quick subsequent time? I believe that’s the level of a caching layer.

“The opposite one is, I don’t suppose that caching must occur outdoors of the cloud knowledge warehouse or the info lake. I believe that you should utilize these methods. The replication of knowledge, in my thoughts, is simply so expensive and so laborious to handle.”

Lastly, for those who had been handed a megaphone and will blast out a message for all the knowledge world, what would you say?


There are a variety of issues in knowledge that you could clear up with expertise, however among the hardest and most essential ones you have to clear up with conversations and other people and alignment and generally whiteboards. Figuring out which form of drawback you’re attempting to resolve at any given time goes that can assist you choose the proper of answer.


I believe the metrics layer is mainly a semantic layer with a further idea of a metric, which is tremendous essential. So I’d simply say, the metrics layer ought to be backed by a general-purpose semantic layer. The spec and the definition of that semantic layer and the abstractions is so unbelievably essential.

Aspect word: I’m personally tremendous enthusiastic about how a metrics layer can work together with an lively metadata platform to supercharge data administration for knowledge groups. It’s been tremendous thrilling to see the metrics layer develop into extra mainstream, which was a prediction I’d made firstly of this 12 months.

Study extra in regards to the metrics layer and my six large concepts within the knowledge world this 12 months.

Report: The Way forward for the Fashionable Knowledge Stack in 2022

Obtain right here →



Most Popular

Recent Comments