Skip to content

Founder notes · 1 min read

DataExpert Business Report: $1.4M in a Year While Subscriptions Shrank by a Fifth

DataExpert.io made $1,397,414 in twelve months, Stripe-verified, with an MRR of $52,398. About half comes from $3,000 live boot camps, and the money arrives in waves.

Oddig business report cover: DataExpert, MRR $52,398, 823 active subscriptions, $4,834,144 all time
← All articles
ODDIG · Business report · DataExpert
Report v1 · Data as of 2026-11-10
DataExpertdataexpert.io ↗ · data engineering courses

DataExpert made $1.4M in a year while its subscriptions shrank by a fifth

A data engineering school run by a creator with a large following: Stripe-verified revenue that arrives in waves around each live boot camp, and a subscription line that is slowly falling underneath.

dataexpert.io
dataexpert.io homepage, captured Nov 11, 2026
Homepage · captured 2026-11-11

Product at a glance

What it isOnline courses in data engineering and AI: live boot camps of about five weeks, a self-paced library, and mock interviews.
Who it’s forWorking engineers and analysts who want to move up, in the founder’s words “like a big tech engineer”.
How it charges$3,000 one time for a live boot camp. $96 a month or $1,150 a year for the self-paced library.
Business modelFree content builds the audience, dated cohorts bring the money, a subscription fills the gaps.
Revenue, last 12 months
$1,397,414
$4,834,144 all time
Fact
Revenue, last 30 days
$104,582
2.0× MRR of $52,398
Derived
Active subscriptions
823
1,045 a year ago, down 21.2%
Derived
Best month ÷ worst
5.1×
$251,176 against $49,281
Derived
Revenue verified via Stripe on TrustMRR · synced 2026-11-10 20:02 UTCResearched 2026-11-11 · Oddig
ODDIG Business Report · DataExpert · v1Data as of 2026-11-10 · Researched 2026-11-11
Quick read shows the scorecard, key findings, what to copy, our predictions and the bear case. Switch to Full report for charts, the price history, the channels and what customers say.

Summary

What this business is

DataExpert.io is an online school for data engineering and AI, founded by Zach Wilson, who says he led data teams at Airbnb, Facebook and Netflix. It gives a great deal away: an open course handbook on GitHub, a free six-week program and a newsletter. It earns from live boot camps at $3,000 each and from a subscription to its recorded library. Revenue is large and uneven, because most of the big months are the ones in which a cohort is enrolling. The method is easy to describe. The scale rests on an audience of several hundred thousand people that took years to build.

FormLive cohorts and a course library
TeamFounder plus six listed instructors
Entry priceFree, then $96/mo
Founded2023-03-22
Newsletter173,000+ subscribers
Reviews4.9 from 669
ODDIG scorecard1 = weak · 5 = strong · our judgement, no total score
Revenue qualityHow much recurs
2/5
Channel spreadNot one channel
4/5
CopyabilityFor a solo founder
2/5
DefensibilityHard to copy the product
3/5
Evidence strengthHow sure we are
4/5
Share

Why these scores? See the reasons in Key takeaways →

Momentum

Mixed. Revenue: last three complete months +25% on the three before; −17% on the same months half a year earlier; MRR −14% since Apr 3, 2026.Derived

Best suited for

Creator (main). The business runs on an audience built by years of public teaching.

Marketer (second). Each cohort is a launch with a date, a price and a deadline.

What you need to start. Real expertise in a field people pay to enter, the willingness to give most of it away, and time. The audience came first.Our read

Key findings

01About half the money comes from cohorts, not subscriptionsThe last 30 days brought 2.0× MRR. Annualised, subscriptions cover roughly 45% to 52% of the last twelve months of revenue.DerivedSee the revenue →
02Subscriptions fell by a fifth while revenue kept comingActive subscriptions went from 1,045 to 823 in a year. All-time revenue grew by $1,371,895 over the same period.DerivedSee both lines →
03The free material comes firstAn open handbook with 45,084 GitHub stars, a newsletter with more than 173,000 subscribers and a free six-week program sit in front of the paid programs.Our readSee the channels →
Evidence tagsFactfrom a listed public sourceDerivedour arithmetic on factsFounderclaimed by the founder or siteOur readour judgement
01

Product overview

A school with three doors: a free one, a monthly one and a $3,000 one with a start date.

The site sells three things. Live boot camps run for about five weeks with classes two evenings a week; the one now enrolling is the Senior AI Analytics Engineering Boot Camp, Nov 16 to Dec 18, 2026, at $3,000. The self-paced subscription opens the recorded library for $96 a month or $1,150 a year. Mock interviews start at $50, and some are free[2].

What the site looks like

dataexpert.io
dataexpert.io homepage on desktop, captured Nov 11, 2026
dataexpert.io homepage on a phone, captured Nov 11, 2026
dataexpert.io/pricing
dataexpert.io pricing page on desktop, captured Nov 11, 2026
dataexpert.io pricing page on a phone, captured Nov 11, 2026
github.com/DataExpert-io
DataExpert-io data-engineer-handbook repository on GitHub on desktop, captured Nov 11, 2026
DataExpert-io data-engineer-handbook repository on GitHub on a phone, captured Nov 11, 2026

Captured Nov 11, 2026 · Screenshots: dataexpert.io and github.com

What to look for

  1. 1
    The promise is a level, not a topic: “Master data engineering and AI like a big tech engineer.”Fact
  2. 2
    The next live boot camp sits on the page with its dates and a “Now Enrolling” label.Fact
  3. 3
    Employer logos stand in for outcomes: OpenAI, Amazon, Meta, Netflix and others.Fact

What to look for

  1. 1
    The pricing page shows one thing: the next boot camp, $3,000, with a start and end date.Fact
  2. 2
    A team option offers “Save up to 25%”.Fact
  3. 3
    The subscription is not on this page. The page exists to fill the cohort.Our read

What to look for

  1. 1
    The star and fork counts sit at the top of the page, where every visitor sees them.Fact
  2. 2
    The repository is a reading list and course materials, not software.Fact
  3. 3
    It carries the school’s name, so each star is also a mention. See §04.Our read

More than one teacher

The homepage lists six instructors, the founder among them, and two guest experts[2]. The founder’s program pages say he led data and software teams at Airbnb, Facebook and Netflix[2]. TrustMRR gives the founding date as Mar 22, 2023; the domain has archived copies going back to December 2021[1][4].

FreeSubscriptionLive boot camp
Price$0$96 a month or $1,150 a year$3,000 one time
What you getA six-week program, newsletter, SQL practice, the GitHub handbookThe recorded course libraryAbout five weeks of live classes with instructors
WhenAny timeAny timeFixed dates

From the live site, read Nov 11, 2026[2]; the free program from the newsletter post that announced it in June 2025[6].

Our readThe free tier is not a trial of the paid product. It is a full course in its own right, and it does the work an ad budget would do elsewhere.Our read
02

Revenue analysis

Two businesses on one Stripe account: a subscription that is slowly shrinking, and a launch business that pays in bursts.

Nov 7, 2026
$18,064
Oct 12Oct 19Oct 26Nov 2Nov 9

Hover or tap a bar, or pick a window. Verified daily revenue, Oct 12 – Nov 9, 2026 (UTC), from TrustMRR[1]. Blue: the twelve days from Oct 29. The next boot camp starts Nov 16[2]. Nov 10 is left out because the day was not complete at sync.

Share
Oct 12–28, per day
$1,431

17 days, $24,328 in total.Derived

Oct 29–Nov 9, per day
$6,539

12 days, $78,466 in total.Derived

Share of 29 days
76.3%

Taken in those 12 days.Derived

Revenue ÷ MRR
2.0×

$104,582 in 30 days against $52,398 MRR.Derived

What the window showsThe next boot camp starts on Nov 16[2]. From Oct 29 the daily average is 4.6 times what it was in the 17 days before. We cannot see which payments are boot camp seats, but the timing and the size of the days fit an enrolment period.Our read

Monthly revenue: three peaks and long valleys

Dec 2025
$251,176
$210K$140K$70K
DecJanFebMarAprMayJunJulAugSepOctNov

Hover or tap a bar. Monthly verified revenue from TrustMRR[1], December 2025 to November 2026. November covers Nov 1–10.

Revenue quality
Revenue ÷ MRR
2.00×
MRR per subscription
$63.67
Typical monthly swing
50% a month

Our arithmetic on the verified data in this report. Revenue ÷ MRR uses the last 30 days; the share of all-time revenue counts the months in the chart above; the swing is the median change between complete months.

December 2025 brought $251,176 and September 2026 brought $49,281, a ratio of 5.1 to 1. The three largest months, December, August and March, hold 47.7% of the eleven complete months[1]. In December 2025 the homepage was selling a 15-week program at $7,497 that started on Jan 2, and two boot camps at $3,000 that started in January and February[4]. The peak and those launches fall in the same weeks; we cannot show which program brought which dollars.Our read

The subscription line

DateMRRActive subscriptionsAll-time revenue
Nov 20, 2025$57,9141,045$3,462,249
Apr 3, 2026$60,899920$4,214,753
Jun 21, 2026$57,694914$4,393,277
Sep 4, 2026$57,708911$4,674,448
Nov 2, 2026$53,450834$4,782,554
Nov 10, 2026$52,398823$4,834,144

Earlier rows from archived copies of the TrustMRR page[3]; the last row from the live page[1]. The November 2025 MRR was labelled an estimate.

Subscriptions, one year
−21.2%

1,045 to 823, a loss of 222.Derived

MRR, one year
−9.5%

$57,914 to $52,398.Derived

Revenue, same period
+$1,371,895

Added to the all-time total.Derived

Subscription share
45–52%

MRR × 12 at its lowest and highest, against 12-month revenue.Derived

MRR tells you about one of the two businesses. Multiply today’s MRR by twelve and you get $628,776, or 45% of the last twelve months of revenue; use April’s higher MRR and you get 52%. Either way, roughly half of what DataExpert collected is not subscription money[1][3].Derived

Each subscription is worth about $64 a month, under the $96 list price. Annual plans, older prices and discount codes are mixed into that average and we cannot separate them.Our read

Market

Paid courses for working engineers are a crowded field. We found no comparable school with named, verified revenue, so revenue is not compared.Our read

03

Pricing: cohorts up, subscription down

Over three years the live boot camp became more expensive and the subscription became cheaper.

ArchivedLive boot campSelf-paced
Jan 2024$2,000 for both tracks; $1,650 for one$1,750 one time
Jun 2024—$1,750 one time. “Bought by 127 students!”
Dec 2024“30% off! Less than 15 seats left in Jan 2025 boot camp!”—
Dec 2025$3,000 for five weeks; $7,497 for a 15-week program$200 a month or $1,500 a year; $100 a month without cloud access
Nov 2026$3,000 for five weeks$96 a month or $1,150 a year

From archived copies of the homepage[4] and the live site[2]. A dash means that capture did not show a price for that product.

Live boot camp
+50%

$2,000 in January 2024 to $3,000 now.Derived

Monthly subscription
−52%

$200 in December 2025 to $96 now.Derived

The self-paced course started as a one-time purchase at $1,750 and became a subscription. The TrustMRR profile still gives the price as $119 a month, which matches neither the December 2025 archive nor today’s site[1][2][4].

What the two moves sayA live seat is scarce: there are only so many evenings and so many people an instructor can answer. A recording is not. Raising the price of the scarce thing and lowering the price of the plentiful one is consistent, and it matches where the money now comes from.Our read
04

Where customers come from

From an audience the founder built by teaching in public, on several platforms at once.

Owned audiencecounts shown on each platform, Nov 11, 2026swipe →
LinkedIn
526K followers

The founder’s personal profile. His headline states a goal of upskilling a million knowledge workers in AI before 2030.[7]

Fact
Newsletter
173,000+ subscribers

On Substack, under the DataExpert.io name. A list the business can reach without a platform’s feed.[6]

Fact
GitHub handbook
45,084 stars

data-engineer-handbook, created November 2023, forked 9,416 times. Course homework and materials are kept here.[5]

Fact
X
51,274 followers

The smallest of the four.[1]

Fact

The GitHub repository is the unusual one. A list of learning resources and the materials for the free program sit in public, where every star and fork is visible to the next visitor. A second handbook, for AI engineers, was created in February 2026 and has 1,258 stars[5].

Conversionwhat turns the audience into buyersswipe →
Free program
$0

A six-week program released free in June 2025, on the school’s own platform and on YouTube. Homework is submitted on the platform.[6]

Fact
Cohort dates
Nov 16 – Dec 18

Every live program has a start and an end, shown on the homepage and the pricing page.[2]

Fact
Discount codes
Up to 25–30%

Code links on the pricing page, and a “30% off” banner in the December 2024 archive.[2][4]

Fact
Team sales
“Save up to 25%”

A buy-for-team option and a discovery call on the boot camp page.[2]

Fact
Our readWe did not check ad libraries for this report. What is visible is reach that was paid for earlier, in years of posts, a free course and an open repository. That is the part a newcomer cannot shortcut.Our read

What we could not read

What the founder posts, and how often, sits behind a login on both LinkedIn and X, so we used only the follower counts. We did not read the YouTube channel. We have no traffic data for the site, whose domain rating is 51[1].

05

What customers say

669 reviews on Trustpilot, almost all five stars, on a profile that invites them.

Reviews
669

Profile claimed in August 2025.Fact

Score
4.9

Out of 5.Fact

Five-star
96%

Four-star 3%, one-star 1%.Fact

Negative reviews answered
100%

As reported by Trustpilot.Fact

ThemeWhat reviewers sayExample
Praise: real-world focusThe course teaches trade-offs such as cost and performance, not only how to use a toolSep 5, 2026
Praise: structureClear lessons and a capstone project with realistic pipelinesJul 7 and Aug 10, 2026
Praise: the teacherThe founder’s own teaching is named as the reason it workedMay 9, 2026
Criticism: paceHard to follow without more background; wanted a slower start before projects3 stars, Mar 3, 2026
Criticism: priceSteep against local salaries in some regions4 stars, May 15, 2026
Criticism: communityWished the community were more activeSep 6, 2026

Themes from reviews on Trustpilot, paraphrased, with the date of one example each[8].

How to read the scoreThe company asks customers to review it, which is normal and also raises the score. Read 4.9 as evidence that the people who answered were satisfied. The three criticisms are more useful: pace, price outside the US, and a quiet community are what a competitor would aim at.Our read
06

Growth strategies

Five moves. The first two build the audience, the next two turn it into revenue, and the last one keeps the proof visible.

P1Open-source the course material
What they didPut a resource list and the boot camp homework in a public GitHub repository.[5]
What happened45,084 stars and 9,416 forks since November 2023.[5]Fact
How you can use itPublish the part of your course that is reference material. People bookmark it, share it and come back to the source.
Cost & conditions You give away content you could have sold. It only works if the paid part is something a file cannot give: feedback, deadlines, a group. 7-day test: publish one useful list or template in public this week and link your course once at the top.
P2Run a free course as the top of the funnel
What they didReleased a six-week program free in June 2025, on its own platform and on YouTube, with homework handed in on the platform.[6]
What happenedNot measurable from outside. The paid programs are all pitched one level up, at people who already have the basics.[2]Our read
How you can use itTeach the beginner level free and charge for the step after it. Free students arrive at the paid door already trained in your method.
Cost & conditions A real course is weeks of work. A thin free course teaches people not to trust the paid one. 7-day test: take your first paid module, make it free for a week, and count how many who finish ask what comes next.
P3Sell by date
What they didGave every live program a start day, an end day and a visible “Now Enrolling” label.[2]
What happenedThe twelve days before the Nov 16 start brought 76.3% of 29 days of revenue.[1]Derived
How you can use itA date gives people a reason to decide. If your product is always available, find something about it that is not.
Cost & conditions Revenue becomes uneven: here the best month was 5.1× the worst. You need cash to cross the valleys. 7-day test: announce a start date for your next group two weeks out and compare sign-ups in the last three days with the first three.
P4Raise the scarce price, lower the plentiful one
What they didMoved the live boot camp from $2,000 to $3,000 and the subscription from $200 to $96 a month.[2][4]
What happenedSubscriptions still fell 21.2% in a year, while total revenue for the period reached $1,371,895.[1][3]Derived
How you can use itCharge most for what you cannot make more of, your live time, and treat recordings as the entry product.
Cost & conditions A cheaper subscription did not stop this one from shrinking. Price alone will not hold subscribers to a course library. 7-day test: list what in your offer is limited by your hours. Price that first.
P5Ask every student for a review
What they didClaimed a Trustpilot profile in August 2025, invites reviews and answers the negative ones.[8]
What happened669 reviews with a 4.9 score, on a site the school does not control.[8]Fact
How you can use itSend the review request at the moment a student finishes something. Reply in public to the unhappy ones.
Cost & conditions Invited reviews read as invited. Leave the critical ones up; they are what makes the rest believable. 7-day test: ask your last ten customers for a public review and answer whatever comes back.
07

Timeline

  • Mar 222023
    ProductFounding date on TrustMRR.[1]
  • Nov 192023
    MarketingThe data-engineer-handbook repository is created on GitHub.[5]
  • Jan 262024
    PricingHomepage: live boot camp $2,000, self-paced course $1,750.[4]
  • Dec 12024
    PricingBanner: 30% off, fewer than 15 seats left for the January 2025 boot camp.[4]
  • Aug2025
    MarketingTrustpilot profile claimed.[8]
  • Nov 202025
    Revenue1,045 active subscriptions; $3,462,249 all time.[3]
  • Dec 32025
    PricingHomepage: a 15-week program at $7,497 and boot camps at $3,000; subscription $200 a month.[4]
  • Dec2025
    Revenue$251,176 for the month, the highest in the last twelve.[1]
  • Feb 162026
    MarketingA second handbook, for AI engineers, is created on GitHub.[5]
  • Aug2026
    Revenue$171,257 for the month.[1]
  • Sep2026
    Revenue$49,281 for the month, the lowest in the last twelve.[1]
  • Nov 102026
    Revenue823 active subscriptions; $4,834,144 all time.[1]
  • Nov 162026
    ProductNext live boot camp starts, at $3,000.[2]
08

Key takeaways

DataExpert shows what a teaching business looks like once the audience exists: generous at the front, expensive at the live end, and uneven month to month.

Why we scored it this way

ODDIG scorecard1 = weak · 5 = strong · our judgement, no total score
Revenue qualityHow much recurs
2/5
The total is large and verified, but about half has to be earned again with each cohort, and active subscriptions fell 21.2% in a year.Derived
Channel spreadNot one channel
4/5
GitHub, a newsletter, LinkedIn, X and a free course. Several channels, and the newsletter list is one the business owns.Our read
CopyabilityFor a solo founder
2/5
The method can be copied by one person. The results depend on an audience of hundreds of thousands, which cannot be copied quickly.Our read
DefensibilityHard to copy the product
3/5
Much of the course material is public. What is hard to copy is the founder’s name and reach.Our read
Evidence strengthHow sure we are
4/5
Revenue, price history and audience counts are first-hand. We do not have enrolment numbers per cohort.Our read
Share

Copy

  1. Put your course material in public where it can be starred, forked and found. P1
  2. Sell by date. A cohort with a start day gives people a reason to decide now. P3
  3. Ask every student for a review on a platform you do not control. P5

Avoid

  1. Reading MRR as the size of a course business. Here it is about half. §02
  2. Planning on launch months. The best month was 5.1× the worst. §02
  3. Expecting this scale without the audience. The follower counts came before the revenue. §04
09

Our predictions

Three claims with dates. We mark each one right or wrong on its date and keep the result on the page.

#PredictionCheck onStatus
#1
November verified revenue passes $100,000TrustMRR monthly revenue. Nov 1–10 brought $57,438 and a boot camp starts Nov 16.
Dec 1, 2026Open
#2
Active subscriptions fall below 800TrustMRR active subscriptions, 823 on Nov 10, down from 834 eight days earlier.
Jan 31, 2027Open
#3
December 2026 or January 2027 brings a month above $150,000TrustMRR monthly revenue. The same two months a year earlier were $251,176 and $148,661.
Feb 1, 2027Open

Track record: the predictions in our earlier reports are still open; the first checks land Dec 1, 2026. We mark each prediction right or wrong on its date and keep the result on the page.

10

What could prove us wrong

Things that would change our read, and how to check them.

  1. The falling subscription count may be a choice. The monthly price went from $200 to $96 and the site now leads with boot camps. The business may be steering people to cohorts on purpose.Check: whether the subscription returns to the pricing page or the homepage’s first screen.
  2. Our split between cohorts and subscriptions is an estimate. We annualise MRR at two points and compare it with twelve months of revenue. Annual plans, refunds and team deals could move it either way.Check: not public. Only the founder has revenue by product.
  3. The reviews are invited. A 4.9 score from 669 reviews on a profile that asks customers to post is a floor for satisfaction among people who replied, not a sample of everyone who paid.Check: Trustpilot marks invited reviews on each entry.
  4. The business leans on one person’s reach. If the founder’s accounts lose reach or he steps back, every channel here is affected at once.Check: newsletter and follower counts at the next review.

Appendix

FAQ

How much does DataExpert make?

$104,582 in the 30 days to Nov 10, 2026, $1,397,414 in the last twelve months and $4,834,144 all time, verified through Stripe on TrustMRR. MRR is $52,398 from 823 active subscriptions.

What does DataExpert cost?

The live boot camp starting Nov 16, 2026 is $3,000 as a one-time payment. The self-paced subscription is $96 a month or $1,150 a year. Mock interviews start at $50, and a six-week program is free.

Why is DataExpert’s revenue so much higher than its MRR?

Because about half of it is one-time payments for live boot camps, which MRR does not count. Revenue rises in the weeks before a cohort starts and falls back afterwards.

Who runs DataExpert?

Zach Wilson, the founder, who teaches alongside six instructors listed on the site. He has about 526,000 followers on LinkedIn and a newsletter with more than 173,000 subscribers.

Sources

  1. [1]
    TrustMRR: DataExpert / TechCreator — revenue, MRR, subscriptions, daily and monthly data. Verified through a Stripe API key. Synced 2026-11-10 20:02 UTC.
  2. [2]
    dataexpert.io — homepage, pricing page and program pages. Read 2026-11-11.
  3. [3]
    Wayback Machine captures of the TrustMRR page for DataExpert on Nov 20, 2025 and Apr 3, Jun 21, Sep 4 and Nov 2, 2026 — MRR, subscriptions and all-time revenue on those dates. Read 2026-11-11.
  4. [4]
    Wayback Machine captures of dataexpert.io — seven homepage captures from Jan 26, 2024 to Oct 8, 2026: programs and prices. Read 2026-11-11.
  5. [5]
    GitHub: DataExpert-io — repositories, stars, forks and creation dates. Read 2026-11-11.
  6. [6]
    DataExpert.io Newsletter on Substack — subscriber count as shown on the page. Read 2026-11-11.
  7. [7]
    Zach Wilson on LinkedIn — follower count as shown on the public profile. Read 2026-11-11.
  8. [8]
    Trustpilot: dataexpert.io — review count, score, star split and review text. The company invites reviews. Read 2026-11-11.

Screenshots © DataExpert.io and GitHub, used for commentary. Oddig has no business relationship with DataExpert.io. Review themes are paraphrased from public reviews. Numbers change daily; check the sources for the latest.

The projects worth a look, in your inbox.

New reports like this one, what is for sale, and who is holding up or fading.

Free. Unsubscribe any time.

The discussion

Comments are closed for this article.

Still building?

Get the next useful signal in your inbox.