Blog/Buying video

How to choose a video company for developer tutorials

Seven things to check, who else could make your tutorials, what moves the price, and a scorecard you can use on your own shortlist.

Kevin Blanco··11 min read·Evaluation

The short answer

Choose a video company for developer tutorials the way you'd hire an engineer: test how they work. Ask whether they install your product before scripting, who checks the code on screen, how they handle API changes and what they measure. Then pay for one tutorial before you sign a series.

01

A tutorial that teaches the old version

Picture a tutorial for a command-line tool. It looks great: smooth zooms, clean audio, a calm voice. At 2:10 the presenter runs a flag that was removed two releases ago, and the terminal output is clearly from somewhere else. Nobody at the company noticed, because nobody at the company was asked to watch it before it went out.

The first comment is a developer saying the command doesn't work. Every viewer after that reads it before they press play.

That's the risk in this purchase. The usual advice for hiring a video company is about portfolios, budgets and how well they listen. It's fine advice. It also misses the thing that sinks developer tutorials, which is whether the steps on screen are true.

02

Why developer tutorials need a different kind of company

A developer tutorial has one job: get someone to a working result. The viewer follows along with a terminal open. If a command fails, they don't blame the editing. They decide the product is harder than it looked, or that the company doesn't use its own product. Both cost you.

A generalist company can make the video look excellent and still get this wrong, because nothing in a typical video process catches a wrong flag. Your engineers end up catching it, usually late, usually by rewriting the script.

So the question to ask isn't whether a company makes good video. It's what in their process stops a wrong command from reaching your audience.

03

Seven things to check

Each row is something you can test on a sales call or in a sample, and each has a red flag you can hear.

What to look forHow to check itThe red flag
They use your product before they scriptAsk to see the notes from the first-run session on a past project"Send us your script and we'll produce it"
An engineer's review is built inAsk who checks the code on screen, at which stage, and who fixes an error found after publishingNobody on their side can read the code
Code on screen is real, readable and runnableWatch a past tutorial on a phone. Can you read the terminal? Is there a link to the repo?Mocked terminals, tiny type, no repo
The craft explains somethingLook for motion that shows what you couldn't otherwise see, like a request being routedEffects with nothing to explain, or an avatar doing the talking
It's built to be foundAsk what ships with the video besides the file: title, chapters, transcript, docs embed, ShortsOne MP4 and nothing else
It's measured on what mattersAsk what baseline and metric you'd agree before production, and when you'd see numbersPromises of views, or reports that stop at views
It's built to lastAsk who owns the source files and how they'd update one chapter after an API changeThe files stay with them, and every change is a new project

04

Why each one matters

  • 01

    Product first

    A script written from your docs repeats whatever the docs got wrong. A script written after someone has installed the product, broken it and fixed it reflects what a new user actually hits.

  • 02

    Engineer review

    Someone has to be able to read the commands. The best arrangement puts a person on the production side who can, and a person on yours who signs off the script and the cut.

  • 03

    Runnable and readable

    If you can't read the code on a phone, most viewers can't either. A repo link turns "I watched it" into "I ran it".

  • 04

    Craft that explains

    Motion earns its place when it makes an abstract idea visible. Anything else is decoration that makes the tutorial longer.

  • 05

    Findable

    Titles, chapters and a clean transcript are how search and AI answers find a video. They're cheap to add when the video is made and tedious to add later.

  • 06

    Measured

    Views tell you a video was seen. Activation tells you it worked. Agree the number before you shoot, so nobody picks the flattering one afterward.

  • 07

    Built to last

    Your API will change. Voice recorded separately from the screen, chapters that stand alone and source files in your hands turn a re-record into an afternoon.

05

Who could make your tutorials

A studio isn't the only answer. Here's the honest range, including when each of the others is the better choice.

OptionStrong atWatch forRight when
A developer-video specialistProduct accuracy, engineer review and craft in one teamSmaller teams, so ask about capacityYou need accurate tutorials and a plan, and don't want to build a team
A generalist video agencyPolish and capacityYour engineers rewrite the scriptsThe video is about the brand more than the product
A freelancerSpeed, low overhead, one person to talk toOne person is a single point of failure, and few cover both code and motionA one-off video with a clear brief
A written-content agency adding videoA network of engineers and a record with developer audiencesVideo may be a newer part of the offer, so ask to watch itYou already work with them on written content
An in-house hireDeep product knowledge, always on handMonths to hire, and one person scripting, recording, editing and packagingYou need a steady, high volume of tutorials
A creator sponsorshipA borrowed audienceYou rent the reach and keep no assetA launch that needs a name attached
AI tutorial toolsThe boring half: transcripts, captions, trimsThey can't verify your product or show effortThey support a process, not replace one

Most teams end up combining two. A specialist builds the anchor tutorials and the system, and the in-house team records the rest to that standard.

06

What it costs, and what moves the price

Prices vary a lot, and few companies publish them. Some charge per video, some a monthly subscription, some by the hour. DevReel's rate is $80 to $120 an hour depending on the service, and a project is quoted against a written scope.

A quote only means something once you know what's inside it. Five things move the price.

  • 01

    Technical review

    How much engineer time goes into checking commands, versions and errors, on their side and on yours.

  • 02

    Capture and motion

    A clean screen recording is cheap. Motion graphics that explain a hard idea take design time.

  • 03

    On camera or off

    A presenter adds a shoot, a person to direct, and more editing.

  • 04

    Rounds of review

    Each round costs hours. A fixed number written into the scope keeps the price honest.

  • 05

    Updates

    The cost of re-recording when the product changes, which is easy to forget and often the biggest line over a year.

07

Score the companies you're considering

The weights below are a starting point. Product depth and the accuracy process together carry 45 of the 100 points, because they're what a developer notices first and what's hardest to fix once a video is live. Change them if your situation is different.

Price is kept out of the score on purpose. It's compared only among companies that clear 70, so a low quote can't make up for a gap in accuracy.

The scorecard

Developer tutorial vendor scorecard

Score up to three companies from 0 to 5 on each row. Nothing you enter is saved or sent.

  • Product and technical depth 25

    Did they install your product and script from that session?

  • Accuracy process 20

    Who checks the code on screen, when, and who fixes an error?

  • Craft on screen 15

    Is the code readable on a phone? Does the motion explain something?

  • Findability 10

    Titles, chapters, transcripts, docs embeds, Shorts from the same session?

  • Measurement 10

    A baseline and one metric agreed up front, numbers in two weeks?

  • Maintainability 10

    Source files handed over, modular chapters, a plan for the next API change?

  • Speed and working model 10

    Turnaround, revision rounds, direct access to the people doing the work?

Score, out of 100

0

Not scored

0

Not scored

0

Not scored

Quoted price, per tutorial

Score at least one company. Price is only compared among companies that clear 70, so a cheap quote cannot cover a gap in accuracy.

08

Questions to ask on the call

  • Product

    • Will someone on your team install our product before scripting?
    • Can we see the notes from a past first-run session?
  • Accuracy

    • Who checks the code on screen, and at what stage?
    • What happens when an error gets through after publishing?
  • Craft

    • Can you show a tutorial where the terminal is readable on a phone?
    • How do you decide what to animate and what to show live?
  • Findability

    • What ships with every video besides the file?
  • Measurement

    • What baseline and metric do we agree before production?
    • When do we see the first numbers?
  • Maintenance

    • Who owns the source files?
    • How would you update one chapter after an API change?
  • Working model

    • Who do we talk to day to day?
    • How many revision rounds are included, and how fast do they turn around?

09

Pay for one tutorial first

A reel shows what a company chose to show you. One paid tutorial on your own product shows how they work. It's the cheapest way to find out whether they ask the right questions, whether your engineers trust the script, and whether the numbers move.

Judge the pilot on the same seven rows you used to choose. The timing below is a rule of thumb from our own projects, not a benchmark.

  1. Day 1

    Agree the baseline and one metric

    For example, average views on your last ten tutorials, or signups from YouTube. Write it down.

  2. Days 2 to 10

    Production

    An engineer on your side checks the script and the cut. Revisions come back in hours, not weeks.

  3. Day 14

    First numbers

    Early signals against the baseline: watch time, clicks to docs, first calls to the API.

  4. Day 30

    A readout

    One page: what moved, what didn't, and a recommendation for the series.

FAQ

Questions people ask

Who makes developer marketing and DevRel videos?

Developer-video studios, generalist agencies, written-content agencies that have added video, freelancers, creators and in-house DevRel teams. Our comparison of developer tutorial video companies names the main ones and what each publishes on price.

How long should a developer tutorial be?

Long enough to finish one task. In our work that's usually one to five minutes, with chapters when it runs longer. Treat that as a rule of thumb from our projects, not a benchmark.

Do our engineers need to be on camera?

No. Many of the best tutorials are a screen and a voice. What you need is an engineer reviewing the script and the cut. On camera only makes sense when the person is part of what you're selling.

Can AI tools make developer tutorials?

They help with transcripts, captions and trims. They can't check that your commands run or show a person working through a real error, which is a large part of why developers trust a tutorial. Use them to support a process, not replace one.

Who should own the files?

You. Put it in the contract, including the source files and the rights to cut the video into other formats. Without the source files, every update becomes a new project with the same company.

How do we keep tutorials current when the API changes?

Plan for it before the first recording. Record the voice separately from the screen, show the version on screen, keep chapters short enough to replace on their own, and budget a re-record for each release.

How many rounds of revisions are normal?

Ask for the number in writing, and for the turnaround on each round. Two rounds is common. At DevReel, revisions come back in hours, in a shared Slack channel with the people doing the work.

DevReel

How we do it at DevReel

We run a video company, so judge us with this scorecard too. Every DevReel script starts with someone installing your product, and an engineer on your team signs off the script and the cut.

Cloudinary's videos usually average around 400 views. Ours passed 9,000 in the first week. For ngrok, watch hours tripled and subscribers went from about 450 to over 800 in the first week.

Cloudinary views

9,000+

in the first week, against a usual ~400

ngrok watch hours

3×

subscribers from ~450 to 800+ in the first week

Now the hard part

Bring us the tutorial you'd start with, and we'll walk through how we'd make it.

Kevin Blanco founded DevReel. He has made 500+ developer-tool videos and spoken at 80+ developer conferences.Last updated November 6, 2026.

Keep reading