Second Measure: "How is Chipotle doing"
Updated: Aug 30, 2021
Bloomberg acquired Second Measure in Dec 2020 (undisclosed amount). As part of the acquisition, one wave of employee exits happened in Aug 2020, while others were part of the acquisition and became Bloomberg employees in Dec 2020. Prior to that the company had raised $25M from YC (2015), Jefferies, Goldman Sachs, Citi, Bessemer. As of 2019 the company tracked 5k brands for it's 150 clients. Second Measure is an "alternative data company". This is a category of companies that crunch google trends, credit card, satellite, web traffic, geolocation and other types of non traditional data sets to help companies and investors see inside other companies to get an edge.
How it started: The founders Michael and Lillian met at EA Sports and managed infrastructure and data pipelines respectively. They constantly helped PMs and game designers with SQL queries to answer questions like "how many gamers made it past Level 1 to Level 2". They ultimately built a self serve tool (Zynga-esque metrics driven) so their own time was freed up. One day a hedge fund ($30B AUM) friend from NYC asked Michael to help load 2TB of data into excel. His only help was an IT guy. As it turned out, this was true for most funds who had a handful of analysts and some back office support. Reminded him of the asks from his gaming PMs; someone not technical asking for behavioral data.
Product strategy: Investors are trying to answer questions like "how is Chipotle doing" in a variety of ways ("how did a food poisoning incident impact revenue", "are same store sales going up", etc), which could all stem from the same data sets. So by talking to potential users, they learned that credit card transactional data is time series, unstructured, messy, and a leading indicator for quarterly performance. The way to answer these types of questions before was with surveys from consulting companies that would charge T$'00s, take weeks and only get back a few hundred data points. Their credit card data, while didn't perfectly track US sentiment, was a good approximation for it, cheaper than the survey option, instantaneous, and had millions of data points. Rather than focus on selling to a persona like investors, they focused on users solving the same problem (Clayton Christensen) at companies and funds. Instead of bespoke insights for one user which would then need to be tested with others, they simply supplied highly cleaned and accurate data. It was up to the user to ask the interesting questions. The editorial team (data scientist, journalist) would put out blogs on the types of questions the data could answer. That type of dog-fooding was not only educational for prospects, but would also help uncover data gaps for product improvement prioritization.
First users and growth: Acquired their first user (a VC) while at YC. All 150 clients as of 2019 were inbound without any outbound sales. Many other VCs followed. Bringing up SecondMeasure dashboards in board meetings would trigger the next CEO to request a demo. Reporters at the WSJ, FT, etc wanted access to insights in return for press mentions.
Product and data operations:
The two products are 1. Pipeline that ingests raw data and spits out useful data 2. Dashboard/ analytics product/ highly specialized Tableau. The dashboard showed data like revenue, LTV, cohorts, company performance, competitive intelligence, benchmarking and consumer behavior. Empowerment is the core purpose, one off research projects handled on a case by case basis.
Messy data. 5k merchants generate 50B data points (unclear time period) with 1B unique transaction codes. Macy's has 3M transaction codes. "Uber SF" is tagged on transactions that happen in NYC. "United" is tagged on airline and United Groceries (Brooklyn) transactions. The two layers of perturbation are the human layer setting up checkout, and data ETL processes.
Some of the questions answered:
"How much more traffic did Uber do than Lyft this last month"
"Is StitchFix cannibalising department store sales"
"Is Peloton outpacing Soul Cycle"
"Is Amazon increasingly relying on Prime subscriptions for revenue"
Limitations: B2B sales (General Mills selling to grocery stores) and product preferences (Uber Limo vs Express) don't appear in the data.
From product and data team members:
PM: Led our core product, selling across the buy-side to hedge funds, private equity, and venture capital. I've also worked on our market research product for corporates, and our data platform. Championed novel features from discovery to launch, increasing revenue by 37%, and usage by 45% in 1 year. Established product standards for the Second Measure coverage list, allowing us to expand it by 120% over 18 months. Built effective monitoring systems to catch errors in our data, reducing defects by 60%. Facilitated partnerships with other data providers, to acquire training data and improve the quality of our platform.
Data Ops: Built a self-service analytics platform and programmatic delivery mechanism to help institutional investor and corporate clients get real-time insights on company performance and consumer behavior.
CBO: Led strategy efforts to build product roadmap and resource plan. Launched major overhaul of analytics front-end.
Dir Product Marketing: All aspects of marketing, including internal and external brand strategy, website launch, content strategy and creation, thought leadership, sales enablement, creative briefing, and PR/creative agency management.
Business Development: Led a multidisciplinary team of four across product, design, and sales to develop the go to market strategy for expansion into Corporate markets. Launched initial pilot product providing retailers with a better understanding of their customers overall wallet spend to better align marketing resources, rolled out product to two new clients. Partnered with marketing to develop and communicate new messaging, value prop and collateral.Hired and managed the first two sales reps and account managers focused entirely on developing the expansion of the market to corporate clients. Developed and managed initial budget and weekly updates to CEO.
Director Of Product Design: Hire horizontal product designers with skills spanning from UX research to Interaction Design. Setting up tooling, managing version control, seeding a design system which they fleshed out, and overseeing the build-out of a code library based on the same. Formalize the UX research process. Special projects such as the bring-up of the Data Products agile team, a full assessment of how our products do and don’t meet client needs, and the research into strategically significant new markets.
Content Marketer: Manage cross-functional content marketing activities to tell data-driven stories. Develop and execute content project plans to drive thought leadership, brand awareness, and to enable sales efforts.
Sr Product Designer: Second designer, but first design individual contributor, hired to help define the user experience for this big data and financial tech startup. As a key contributor to multiple shipped products, I solved challenging data visualization and user workflow issues while collaborating with engineers, data scientists, customer engagement managers, and product managers.
Design Lead: Designed over 5 features on an enterprise B2B web app to improve data visualisation, usability and user experience. Planned, conducted, and documented remote user research sessions. Collaborated cross-functionally on-site & remotely to improve operational processes and user interfaces.
GTM: From team members:
CBO: Led Revenue, Finance, BD, & Ops as headcount grew 4X. My team built pipeline tracking, forecasting, budgeting, and planning systems.
Head of Client Engagement: Built out Second Measure's strategic account management and customer success functions as a founding member of the revenue team.
Director Of Business Development: Created & lead execution of go-to-market revenue strategy to scale corporate brand line of business for both new logo acquisition & expansion of existing corporate client base spanning growth-stage companies to Fortune 500. Closed 30+ new corporate logos (1.5 yr period).
Director of Business Development: Sourcing new revenue opportunities for a highly complex alternative data SaaS product which allows investors to generate alpha and manage risk. Engage with institutional investors across the East Coast (hedge funds- quantitative and fundamental, asset managers, investment advisors, venture capital and private equity firms) across all stakeholder levels. Performed market-sizing for addressable client base, analysed holdings and investment mandates to tailor strategic approaches to engage stakeholders throughout a consultative sales process. Collaborate across verticals internally to improve UX and enhance product development, engaging with applied data, product, engineering and client services to communicate gaps in client workflow.
Engineering and Data Science: From team members:
SWE: Built a pipeline orchestration system using AWS and Kubernetes which improved reliability and speed of customer data delivery. Implemented data corrections and transformations in Spark and Scala that enabled new insights by both the data team and customers.Deployed and managed services in Kubernetes using Helm, facilitating simpler, more reliable deployments and releases. Managed infrastructure with Terraform, automating repetitive tasks and increasing the recoverability and resilience of systems. Automated legacy code to enable regular updates of internal tooling vital to the data scientists' work.
Data Scientist: Implemented client-facing forecasting for financial metrics in both short and long time series contexts. Led research and implementation of comparable-store sales feature, a first in the transactional data space.
Data Scientist: Forecasting, Building models, Machine Learning, and Building Data Products. Mostly on the Machine Learning/Statistics Pipelines, Time Series, and Alerting end of the spectrum. Technologies: Python (lots of libraries), Data Pipelines, Spark, SQL, Key/Value stores.
SWE: Engineer on the Foundational Data Team. As a member of the Foundational Data engineering team, I was responsible for designing and implementing reliable data pipelines for processing transaction data. Implemented critical technical requirements for processing data, including implementing distributed data query layer using Presto and Hive Metastore, as well as platform and orchestration using Kubernetes and Argo Workflows. Implemented a monitoring solution for our pipeline orchestration in GO. Added features in Spark (scala) to critical parts of our main ETL pipeline in close collaboration with Data Scientists, such as Intermediary Tagging. Core developer for Second Measure client data delivery services, including client web application and data feeds. I was responsible for implementing customer facing features for data analysis, such as Quarter to Date, and building internal management tooling around our data delivery services. Designed and implemented critical authentication and authorization functionality in client application, including multi-factor authentication and the migration of the auth microservice. Added critical customer and internal facing features, such as the Quarter-to-Date V2 page and the data-feed management page respectively. Established software patterns such as parameter sanitation and validation methods, and enhancements to foundational libraries.
SWE: TypeScript, React, Redux, Cypress
Data Science Intern: As part of the Data Products team, I developed one of the first predictive models to predict per-company quarter-end revenue from observed sales. This project has since made it into the platform to augment the client's investment decisions with dependable forecasts for revenue. As part of the Client Services team, I created a framework and metrics to measure the engagement of a company's customer base based on power user curves. This framework provides potential investors with a way to quantitatively measure how often customers purchase from companies.
Team: As of 2019, 60 people. 20 PhDs. 50-50 engineers and data science. Data science team backgrounds vary from genetics, neuroscience, climate science, string theory, etc, but the common thread is statistics and first principles (if given a data set, does one over trust it and dive into charting or diligently look for dragons in the data and state assumptions before charting).
From team members:
Head of talent: Reporting to the founders, was first and sole recruiting leader brought in to develop hiring processes and scale team. Hired over 80 employees (exec/leadership to individual contributors in SF Bay Area and NYC offices) across marketing, sales/client engagement, data science, engineering, finance, legal, people, operations, product, and design. Through transparent recruiting approach with candidates, attained a 78% offer acceptance rate while competing with the largest tech/finance companies in the SF Bay Area and NYC. Forbes Best Startup Employers 2020, BuiltIn Best Small Companies to Work For, and achieved a Glassdoor employee rating of 4.8/5.