The numbers Brief Episode map Timeline Takeaways Charts Quotes Clips Glossary Go next Transcript

Episode 133 · Enterprise · 43 min

Ask the diner, not the critic

G2's wager is that the person eating the food knows more than the restaurant critic who can only interview the diners. So the verdict on business software should come from verified users, free to publish and free to read. Behind it sits a founder who once waited nine years to get his own company into an analyst report, and a catalogue of 2,100 categories now carrying 2.8 million reviews.

SaaS & enterprise Founder journey Marketing & brand

Godard Abel and Sarah Rosier

Co-founder & CEO; Chief Product Officer, G2 · with Vishal Krishna

42:52

Said in this episode

▶ 9:22

2,100

Software categories on G2

The taxonomy the research team maintains; product counts were given as 180,000 early on and 'over 100,000 listings' later in the same conversation, so treat the catalogue size as approximate.

▶ 27:15

2.8M

Reviews on the platform

Stated by the chief product officer while describing what AI summarisation has to digest; roughly 10 million visits a month sit on top of it.

▶ 27:24

~10M/mo

People coming to G2 each month

Offered in response to the host's guess of four million users: 'about 10 million a month that are coming to G2'.

▶ 15:11

9 → 12 yrs

Time to a report, then to 'leader'

Abel's first company waited nine years to appear in an analyst report and twelve to be named a leader. That is the wait he built G2 to abolish.

▶ 16:00

$50–100k

Cost of marketing a quadrant placement

Abel's figure for what a vendor pays for the right to promote its analyst placement, quoted as 'at least' that range.

▶ 34:07

$400M

Oracle's price for BigMachines

After never raising beyond the original round and growing past $50 million ARR; the figure is Abel's recollection on air.

The brief

The argument in sixty seconds

Abel's claim is that the software analyst business was built for the wrong customer. The old reports covered only the biggest vendors and reached mainly Fortune 500 CIOs with subscriptions; his first company, BigMachines, took nine years to get into a report and twelve to be called a leader, and he says vendors still pay $50,000 to $100,000 simply for the right to market a quadrant. G2's counter-model borrows from Amazon and Yelp (Zomato, in the host's translation), but repairs the thing that broke consumer reviews: reviewers are verified through LinkedIn, so a buyer can see whether the person praising a CRM actually does that job, in that function, in that city. The analysts do not disappear; they change job. G2's research team writes no opinions, only the taxonomy: 2,100 categories and their inclusion criteria, adding vector databases the moment the market demanded one. The opinions come from users, in real time, which is why listing is free. That free listing is how Freshworks and Chargebee let customers vouch for them to American buyers who had no reason to trust an Indian SaaS vendor. Now the catalogue is itself the problem: Abel calls G2 a supermarket with 2,100 aisles, so Monty, an agent built on OpenAI's APIs and sitting on 2.8 million reviews, is being taught to match a described business problem to a product. Underneath all of it runs a founder's argument that fortitude beats smarts: $19 million of a $20 million round burned in three years, a company cut from 70 people to 20, and seven years before it worked.

Worth your time if you are

Small-business buyers who pick software by brand name

Indian SaaS founders selling into a US market that has never heard of them

Product leaders whose catalogue has outgrown its own search box

Founders three years in with one year of runway left

Engineers weighing a capability centre against a full-stack product office

Episode map

Where the conversation travels

Every block is a chapter, coloured by what it's about. Click any of it to jump straight to that minute on YouTube.

SaaS & enterpriseFounder journeyMarketing & brandAI & machine learningHiring & talentGCCs & services

01Cold open: a marketplace built on reviews 0:00 Vishal sets the premise (small businesses and procurement teams buy by brand name and never look at reviews), then introduces G2 as the world's largest software marketplace and its CEO and chief product officer as his guests. 02A $100 lunch versus $100,000 software 2:17 The founding irritation: you can research a hundred-dollar lunch better than a hundred-thousand-dollar CRM, because analyst reports covered only the biggest vendors and only Fortune 500 CIOs held subscriptions. 03The critic who cannot eat the food 3:57 With software eating the world and every function getting a purpose-built app, Abel argues the analyst is a restaurant critic barred from the kitchen, interviewing diners instead of tasting, while buyers fall back on the one app a friend happens to know. 04Trust starts with a LinkedIn profile 7:08 Anonymous or paid reviews are what broke consumer trust, so G2's first rule was identity: verify the reviewer through LinkedIn, which also supplies the context (function, seniority, city) that lets a buyer find someone like themselves. 05Categories, the flywheel and a thank-you 8:52 Across 2,100 categories the research team decides what a category even is (vector databases became one), while targeted outreach and a thank-you unrelated to the star rating seed the first reviews in thin B2B niches. 06Freshworks, SaaSBOOMi and the free listing 12:45 Any company can list free and generate reviews; Abel credits that route for Indian unicorns like Freshworks, whose founder decided customers would vouch for him better than he could, and points to a Bangalore G2 Live with 1,200 SaaS people signed up. 07Gideon Gartner's eight binders 15:30 Asked to draw the line against the incumbent, Abel recalls pitching the idea to Gartner's own founder, a fellow MIT alumnus two decades out of his company, who agreed it had stopped innovating and tried to hire him instead of investing. 08The CIO no longer buys the software 18:23 Even inside giant corporations the buyer is now a project manager doing research, and the modern path runs bottom-up: Salesforce and Datadog started with SMBs and individual practitioners before moving upmarket. 09180 people, Ruby on Rails, Snowflake 19:50 In a war for talent against every new capability centre, G2's India team of roughly 180 runs engineering, operations and market research out of an office opened that week, on a Rails stack with its data cloud on Snowflake. 10Monty and the 2,100-aisle supermarket 21:53 Choice became the bottleneck. Abel likens the catalogue to a supermarket with 2,100 aisles, so Monty takes a described business problem and matches it to products, with a market-intelligence agent aimed at vendors who cannot see why buyers prefer a rival. 11India's toughest customers call it clunky 23:32 With around 400 customers in the region, service-provider reviews now covering the Wipros and TCSs, and a full-stack office where engineers can walk over to sales, G2 absorbs the verdict that its own back end is not as elegant as Stripe. 12AI is a tool, not a profession 25:56 Monty is built on OpenAI's APIs and Abel doubts prompt engineering is a separate career, since he wants high-aptitude engineers who will figure it out. An all-AI hackathon that week aimed at discovery and at summarising 2.8 million reviews. 13Ten percent India, and the Japan exception 27:40 India is at least a tenth of customers and users and likely more of the reviewers, APAC over a fifth; but APAC is not one market, so Japan got a partner, a separate site and its own language. 14Nineteen million burned, seventy people cut 30:13 BigMachines raised about $20 million on an idea at the tail of the dot-com boom, had $1 million left three years later, shrank from 70 people to 20, and only started working in 2007, before Oracle bought it for $400 million. 15Sleeping under a bridge in Paris 34:35 His father was an industrial-pump entrepreneur who moved the family from Germany, and his son's first customer and first angel. He modelled the emotional fortitude Abel now rates above intelligence as the founder's real differentiator. 16Inner engineering, ice plunges, teenage phones 37:50 Marty Cagan's Transformed gives the two a shared product language, Sadhguru's Inner Engineering gives Abel meditation, ice plunges and the reframe that life is choice rather than something happening to him. Then both admit no law will fix the phone.

Timeline

The arc, briefly

Twenty-five years in enterprise software, and the review site that came out of the waiting.

1980sGartner fits its moment. There was no internet, the reports covered only the biggest vendors, and the readers were Fortune 500 CIOs with subscriptions. When Abel later pitched a modern Gartner to Gideon Gartner in New York, the founder had been out of his own company for about 20 years and agreed it had stopped innovating. He had eight binders of a new idea with him and would not let Abel read them.

2000BigMachines raises $20 million on an idea. The tail end of the dot-com boom, and software for selling configured machinery online to manufacturers like GE. Winning GE took nine years, getting into an analyst report nine, being called a leader twelve.

2003Three years in, $19 million is gone. One million left, the sales plan missed every year, and nobody willing to put in more. Abel cut the company from 70 people to 20 and gave it a year to get cash-flow positive.

2007The cloud gets popular and the company works. Salesforce and Oracle had taught buyers to trust online software and handed BigMachines distribution as a preferred partner, and it repositioned as CPQ. It passed $50 million of ARR without raising again, and Oracle bought it for about $400 million. The same year Abel started coming to India, to an engineering team in Hyderabad.

2011Freshdesk, and software eating the world. The host dates Girish's company to about 2011, when it was one product rather than a suite. Abel puts G2's own start around the time Marc Andreessen wrote that software was eating the world, meaning a purpose-built app for every business function.

2012-14Every review gets a name on it. Consumer reviews broke because they were anonymous or paid, so G2 verified reviewers through LinkedIn, which works in an industry where everyone has a profile. Listing stayed free, because Abel did not want anyone else waiting nine years.

2020Chennai, just before everything shut. Abel went to SaaSBOOMi and found 700 or 800 SaaS companies already in the room, in a country he had first visited when it was mostly back-office engineering for global firms. Freshworks and Chargebee used free listings to get customers to vouch for them to American buyers.

TodayA supermarket with 2,100 aisles. Over 100,000 listings, 2.8 million reviews and about 10 million visits a month, with Monty built on OpenAI's APIs to turn a described business problem into a shortlist. India is at least 10% of customers and users, the Bengaluru office is around 180 people, and G2 Live runs there for a third year.

Takeaways

Ideas to carry out of this hour

The analyst model priced out the people who actually buy

Abel's grievance is structural, not personal: analyst coverage went to the biggest vendors and analyst subscriptions went to Fortune 500 CIOs, so everyone else bought blind. His own company took nine years to appear in a report and twelve to be named a leader. He says marketing a quadrant placement still costs a vendor at least $50,000 to $100,000. G2's answer is a different gate rather than a better ranking: any company can list free, and validation is earned from customers rather than bought from an analyst.

▶ 3:00 ▶ 15:11 ▶ 16:00

Ask the person eating the food

The traditional analyst, in Abel's image, is a restaurant critic who must stand outside and ask diners how the meal was. G2 inverts that: the reviewer is the person using the software, which he argues gives three compounding advantages. The user knows more than the analyst, the signal is real time rather than annual, and with no primary-research payroll the whole thing moves faster and reaches far more buyers.

▶ 4:56 ▶ 5:10 ▶ 16:28

Trust is an identity problem before it is a content problem

What poisons consumer reviews is anonymity and payment, so G2's first design decision was that you must know who wrote the review, and in tech, LinkedIn already holds that identity. Verification does double duty: it proves a real business professional wrote the words, and it supplies the context that makes them useful, since a buyer wants the reviewer who shares their function, their company size, their city. The product team now also validates in-context that the reviewer is genuinely the user of the software.

▶ 7:22 ▶ 7:51 ▶ 8:21

The taxonomy is the product

G2's research team writes no opinions. Its job is to be steward of a taxonomy of some 2,100 categories, defining each one and the criteria for inclusion, and creating new ones as the market moves, which is how vector databases became a category. In thin B2B niches where there are no millions of users, that structure has to be seeded: targeted outreach using LinkedIn skills to find, say, engineers who actually work with LLMs, and a thank-you for the time that is deliberately decoupled from whether the review is one star or five.

▶ 10:19 ▶ 10:48 ▶ 11:32

Reviews are how an Indian SaaS company borrows trust it hasn't earned

The hardest problem for an Indian SaaS founder selling into the US is that nobody has a reason to believe them. Abel's account of Freshworks is that its founder concluded there was no better answer than letting customers speak for him, and today the company carries thousands of reviews across service, support and sales categories, having started with a single product. Listing free and letting the reviews accumulate is a distribution strategy for anyone who cannot buy their way into a report.

▶ 13:00 ▶ 13:15 ▶ 13:29

Discovery broke under the weight of its own catalogue

Two thousand one hundred categories solved the coverage problem and created a navigation one: Abel calls G2 a supermarket with 2,100 aisles where you no longer know which aisle to shop in. Monty, built on OpenAI's APIs over 2.8 million reviews, is the fix: describe the business problem in prose and let the agent do the matching. The same data pointed the other way becomes market intelligence for vendors, who can see internal feedback but never why buyers prefer a competitor.

▶ 22:21 ▶ 22:34 ▶ 23:02

Emotional fortitude, not intelligence, is the differentiator

Abel is blunt that the founder myth gets the causation wrong: there are plenty of smart entrepreneurs, and what separated the successful ones he knows was emotional perseverance. His evidence is his own: $19 million of a $20 million round gone in three years, a decision to cut from 70 people to 20 and reach cash-flow positive within a year or close, four MIT friends he had recruited and felt he was ruining, and seven years before the company felt like it was working. The tactics and the product cycles have accelerated since; he argues the fortitude required has not changed at all.

▶ 31:11 ▶ 32:08 ▶ 36:20

The numbers, drawn

What the episode measures

Every figure below was said on air, with timestamps included and caveats kept.

How long the old model made you wait

years

All four spans as recounted on air: the company started in 2000 and only began working in 2007; nine years to close GE; nine years into a report and twelve to leader status.▶ 15:11

BigMachines, from near-death to Oracle

$ million

Figures as recalled by Abel in the interview; ARR was described as 'over 50 million' and no further capital was raised after the original round.▶ 34:07

The critic outside, the diner at the table

the two models set against each other, on Abel's own terms

Abel's comparison drawn out. His image is a restaurant critic who has to stand outside and ask diners how the meal was; the counts on the right are G2's own, given later in the same conversation.▶ 4:56

A supermarket with 2,100 aisles

the same corpus, pointed in two directions

Coverage was the first problem and navigation became the second. Abel says you no longer know which aisle to shop in, so the agent does the matching; run the other way, the same reviews tell a vendor what internal feedback never will.▶ 22:21

Worth keeping

Lines that stay

One of our founding ideas was: why can you get more information on a $100 lunch than a $100,000 CRM software?

— Godard Abel ▶ 2:31

My frustration was that it's like a restaurant critic that can't eat the food. They stand outside the restaurant and ask diners how their food was. Why not ask the person eating the food — the person using the software?

— Godard Abel ▶ 4:56

My first company took nine years to get into a report, twelve years to get into a leader. I didn't want anyone to have to wait that long. And I didn't want entrepreneurs to have to pay.

— Godard Abel ▶ 15:11

People are always like, 'Oh, that entrepreneur is successful because they're so smart.' I don't think that's the differentiator. There's a lot of smart entrepreneurs. It's the emotional perseverance.

— Godard Abel ▶ 36:20

You're moving too slow if you're not trying to take some risk and fail.

— Sarah Rosier ▶ 37:24

Clips that travel

Short on time? Start here

Small-business buyers who pick software by brand name

Why a $100 lunch beats a $100,000 CRM

The whole thesis in five minutes: the analyst model's blind spot, nine painful years selling to GE, and the restaurant critic who can't eat the food.

2:17 → 7:08 · 5 min ▶ Watch clip

Marketplace and community builders starting from zero

What makes a B2B review believable

LinkedIn as the trust layer, how a category gets seeded when it has no millions of users, and why the thank-you is decoupled from the star rating.

7:08 → 11:17 · 4 min ▶ Watch clip

Indian SaaS founders selling into a market that has never heard of them

Freshworks, SaaSBOOMi and the free listing

The concrete playbook for borrowing credibility from your own customers when you cannot buy your way into an analyst report.

12:45 → 15:30 · 3 min ▶ Watch clip

Product leaders shipping AI on top of a large catalogue

Monty and the supermarket with 2,100 aisles

Where an agent actually earns its place, matching a described problem to a product, plus India's 400 customers and the customer who called the back end clunky.

21:53 → 25:56 · 4 min ▶ Watch clip

Founders three years in with one year of runway left

Nineteen million burned, seven years of grinding

The unvarnished version: cutting 70 people to 20, feeling he was ruining his college friends' careers, and a father who said he could be happy sleeping under a bridge.

30:13 → 35:10 · 5 min ▶ Watch clip

Glossary

The jargon, unpacked

Analyst quadrant

The grid on which a traditional research firm plots software vendors; Abel says a placement costs a vendor at least $50,000 to $100,000 to market, and took his first company nine years to reach.

Taxonomy

G2's structure of roughly 2,100 categories plus the criteria for what belongs in each, maintained by a research team that supplies the guard rails but never the opinions.

Review flywheel

The loop where seeded early reviews attract buyers, whose traffic attracts vendors, who bring more reviewers. That loop is the thing G2 has to kick-start by hand in every new category.

Social proof

Using verified customer reviews as evidence in a sales conversation, the other half of what vendors take from G2 alongside product feedback.

Consumerization of IT

The expectation that business software feel as immediate as a consumer app, the standard by which an Indian customer judged G2's own back end clunky.

Monty

G2's AI agent, built on OpenAI's APIs over the review corpus, meant to work as a software matchmaker: describe the business problem, get product recommendations.

CPQ (configure, price, quote)

Software that assembles and prices complex product configurations for a sales quote. It was the position BigMachines took beside Salesforce and Oracle, and the product it became after acquisition.

ARR

Annual recurring revenue, the subscription-software yardstick; BigMachines passed $50 million of it before Oracle bought the company.

Connections

If this resonated, go here next

EP 95 · Enterprise

Full transcript

The whole conversation, searchable

Auto-generated captions, lightly cleaned. Click a timestamp to open that moment on YouTube.