Three engineers from Israel's most secretive intelligence unit built a dashboard for marketers. Eleven years later it sits underneath roughly $10 billion of ad spend a year - and nobody outside growth Slack channels knows their name.
It pulls cost data from forty ad networks, joins it against install events from iOS and Android, layers in SKAdNetwork conversion postbacks, attaches creative metadata, and spits out one number: ROI on yesterday's spend for a mobile game with thirty million daily users. Total wall time, roughly four seconds.
The query did not run on the game studio's servers. It ran on Singular's. That is the entire shape of the company. Marketers used to wake up to forty browser tabs; Singular replaced the tabs with one screen. It is, in the words of one customer, "unsexy software people refuse to stop paying for."
The numbers around it are casually large. More than $10 billion in annual ad spend optimized. Over $7 billion in cumulative spend tracked since launch. Two thousand integrated ad and measurement partners. Official partner status with Google, Meta, Apple, X, and Pinterest. Lyft, Yelp, LinkedIn, Match, Twitter, Rovio, Symantec, Zynga. The company has been quietly central to mobile growth marketing since roughly the iPhone 6.
For most of the 2010s, a senior growth marketer's actual day looked like this. Log into Facebook Ads. Export CSV. Log into Google. Export CSV. Log into TikTok. Export. Then AppLovin, Snap, Unity, Liftoff, Vungle, ironSource, IronSource-but-the-renamed-version. Paste into a master sheet. Argue with finance about what "cost" means. Re-do the joins. Pretend the result is decisive.
Multiply by every channel, every creative variant, every geo, every device. The math does not arrive on time. Decisions get made on vibes. Bad campaigns run for three more days because nobody could prove they were bad. That is the problem Singular set itself against - marketing's accountability gap, which is wider and more expensive than anyone politely admits.
A field note: when you stand a CFO and a growth lead next to the same dashboard, the dashboard almost always wins the argument. This is the entire commercial premise.
Gadi Eliashiv, Eran Friedman, and Susan Kuo had spent the early 2010s at Onavo, the Israeli mobile analytics company that Facebook acquired in 2013. Gadi had written his first code at twelve and led a team in Unit 8200, the IDF's elite signals intelligence outfit. Eran had been Onavo's third hire. Susan came from InMobi. They had all watched mobile measurement get inhaled by Menlo Park.
Their bet was almost dull on paper. While everyone else in 2014 was building yet another "mobile measurement partner" with a slightly nicer pie chart, the Singular team argued the real prize was upstream: the ETL layer. If you owned the pipes that pulled in every ad network's cost data, normalized the schema, and married it to the attribution stream, you would own the only screen that ever needed to be open.
Pundits suggested this was unglamorous. The founders agreed. They built it anyway.
CEO. Former Head of Research at Onavo. Unit 8200 alum. Started coding at 12, which is mentioned in every profile because there is no escaping it.
CTO. Onavo's third hire. Built the engineering org that became Facebook's Israel HQ. Now defends a Kafka and Iceberg-shaped opinion about data.
COO. Mobile ad tech veteran via Onavo and InMobi. The reason Singular's customer list reads more like a Fortune 500 attendance sheet than a startup pitch deck.
Seed round of $5M from General Catalyst, Translink Capital, and Method. The pitch is "marketing ETL," which causes investors to ask what ETL is.
Roughly $15M to expand the attribution side of the platform. The first big-name app developers sign annual contracts.
Lyft, Yelp, LinkedIn, Airbnb, Symantec, Zynga, Match, and Twitter are by now all confirmed customers. The dashboard quietly optimizes more ad spend than any of its rivals.
Apple's privacy changes break the deterministic attribution model. Singular ships an SKAN-native solution faster than the field expects, and re-rates a category that briefly looked existential.
Customer-review-based rankings put Singular at #1 in mobile measurement. The company posts the badge and otherwise carries on.
A grounded AI layer that answers marketing questions using only the customer's own data, with sources cited. Refreshingly hallucination-averse.
Strip away the branding and the platform is six things that have to work together or none of them are worth anything.
Pipelines into 2,000+ ad networks. Pulls cost, creative, and performance into a normalized schema. The least glamorous, most load-bearing piece.
MMP across iOS, Android, web, CTV, PC and console. Deterministic, probabilistic, SKAdNetwork, AdAttributionKit. All the alphabet soup.
AI ties individual visuals, hooks, and messages to downstream revenue. Tells you which thumbnail actually paid the bills.
Combines in-app ad monetization data with user-level attribution to produce a true LTV picture - the number most teams quietly want.
Real-time fraud detection across installs, in-app events, and post-install activity. Saves customers more than the contract costs.
Grounded LLM layer trained only on the customer's data. Cites sources. Refuses to invent metrics. A small revolution in MarTech.
Customer logos read like a list of apps you used this morning. Lyft. Yelp. LinkedIn. Match. Airbnb. Symantec. Zynga. Rovio. Twitter, now X. The company sustains itself because growth teams at all of them looked at the cost of churning Singular and concluded that the alternative was forty browser tabs again.
Then there is the partner side. Official measurement partner status with Google, Meta, Apple, X, and Pinterest is a small handful of certifications that the rest of the MMP category collectively dreams about. Snowflake integration ships natively. The stack reads like a data engineer's wishlist: Kafka, Iceberg, Druid, DynamoDB, AlloyDB, Kubernetes.
Ask the founders what they are trying to build and you get a sentence so flat it sounds like a tax form: unify the world's marketing data. Translated, it means: every dollar a company spends on advertising should be traceable to the user it produced, the revenue that user generated, and the outcome that revenue funded.
That sounds like table stakes until you try to do it. Apple's privacy changes shattered deterministic attribution. SKAdNetwork patched the hole imperfectly. AdAttributionKit patched the patch. The ad networks themselves report numbers that are, charitably, optimistic. Doing this work credibly is harder in 2026 than it was in 2018, and Singular is one of a very small number of companies that has not quietly given up trying.
Singular AI is the most recent expression of the bet. It refuses to answer questions it cannot ground in the customer's own warehouse. It cites. It abstains. In a market full of confidently wrong chatbots, that abstention is the feature.
The big tectonic story in marketing isn't AI. It's privacy plus AI. The signal that mobile attribution used to ride on has been progressively wrung out by Apple and Google. At the same time, the volume of channels is exploding - retail media, CTV, AI assistants, audio, in-game. Singular's Q1 2025 trend report noted that ChatGPT had become a top-30 mobile advertiser, a category that did not exist eighteen months earlier.
In a world where the signal is noisier and the channels are more numerous, the dashboard that resolves them into one number becomes more valuable, not less. That is the company Singular is positioned to be. It does not require the future to be utopian. It requires growth marketers to keep wanting to know what worked. They will.
Back in Tel Aviv, the four-second query has returned. Yesterday's spend on the mobile game's iOS campaign produced 14,212 installs at a blended CAC of $3.41, with one creative variant outperforming the others by a 31% margin. The growth lead reads the dashboard, kills two campaigns, doubles a third, and is closing his laptop by 9:14 a.m.
Eleven years ago, that same exercise took two analysts and a working morning, and the answer was probably wrong. That is the difference Singular has made: not a category invented, just a category quietly made functional. Boring is the right word for it. Boring is also, here, a compliment.