The short version

Static (15) and feature callout (20) ship today — no people, no schedule. Testimonial (60) and above need someone on camera, which means a rate, a turnaround and a calendar you do not control. Everything else sits in between.

Put that number next to every gap on your list and sort ascending. The cheapest thing you have never tested is almost always the right next test.

The scores

Sorted cheapest first. Low is 0–30, medium is 31–55, high is 56–100. The score answers one question: what does it cost you to answer an opening in this format?

Format	Difficulty	Band	What it takes
Static	15	Low	A photo you already have, cropped.
Feature callout	20	Low	An afternoon in Figma. No people, no schedule.
Product demo	35	Medium	A tabletop shot or an animated reveal. Hands, not faces.
Voiceover	40	Medium	A script, a voice, and footage you can cut to.
Before/after	55	Medium	Two states, captured consistently. Often needs time to pass.
Testimonial	60	High	A real person, on camera, saying something usable.
Unboxing	60	High	Product shipped to a creator, plus their turnaround.
POV	70	High	Staged first-person camera work throughout.
GRWM	75	High	A creator, a routine, and a location that reads as theirs.
Day-in-life	80	High	A creator, a script, a shoot, and several locations.
Skit	80	High	Casting, a written scene, and direction. It can also just not be funny.
Other	45	Medium	Where an unscored format lands. Mid-band on purpose: score an unknown cheap and it floats to the top of your list on a guess; score it expensive and it never gets made.

Scope. These are Likely's production-difficulty scores, used to rank creative openings by what they cost to answer. They are judgements about production, not measurements of performance — nothing here says a skit outperforms a still. Reuse them under CC BY 4.0.

Why difficulty belongs in the ranking

Rank openings by size alone and every item at the top takes a month to make. The list is accurate. It is also unusable.

Every competitor tool will happily tell you that your category runs a lot of day-in-life content and you run none. Fine. Now what? That gap needs a creator, a brief, a shoot and an edit. If your answer to "what should we make next" is four things that each need a production schedule, you have not made a plan. You have made a wish list.

Difficulty is the axis buyers never see in a competitor tool. Likely ranks openings on three numbers — impact, confidence and difficulty — and difficulty is the one that turns a ranking into a decision.

Score	Answers	Where it comes from
Impact	How much does the category commit here?	Category ad volume in the cell. Volume is the only public proxy for spend, and spend is the only proxy for conviction.
Confidence	Is this real, or is it noise?	How many distinct brands do it, how long they have kept doing it, and how much of it there is.
Difficulty	What does it cost us to answer?	The format the category actually uses in that cell, scored on this page.

The best opening is rarely the biggest one. A medium-impact gap you can answer with a feature callout this week usually beats a high-impact gap that needs a creator booked for next month — because you will have learned something before the second one has even been scheduled.

What the three bands actually mean

The distinction that matters isn't the exact number — it's which of three questions you have to answer: can a designer do this today, does it need a person on camera, or does it need a shoot?

Low (0–30) — a designer can ship it today

No people, no scheduling, no dependency on anyone outside the team. Static and feature callout live here. If your coverage grid shows a gap you can answer in this band, there is genuinely no reason not to test it this week.

Medium (31–55) — editing, a voice, or time

Product demo, voiceover and before/after. You need a bit more than a design tool — a tabletop setup, a script, someone to read it, or in the case of before/after, actual time to pass. Still no casting, still no location.

High (56–100) — a person on camera, or a shoot

Testimonial through skit. The moment a real human has to be filmed saying or doing something, you have inherited a schedule, a rate, usage rights, and a turnaround you don't control. That is a different kind of commitment, and it is why six of the eleven formats sit in this band.

The reasoning behind each score

The scores are ordinal — what matters is the order and the gaps between the bands, not whether a testimonial is exactly 60.

A few of the placements are worth explaining, because they are the ones people push back on.

Feature callout at 20, not 40. It looks like work because it carries a lot of information. It isn't — it's a layout job with no dependencies. This is consistently the most under-used cheap format we see.
Before/after at 55, above voiceover. The production is simple. The consistency is not: two captures, matched lighting and framing, often weeks apart. That gap is where these fall apart.
Unboxing at 60, same as testimonial. Filming is trivial. Getting the product to a creator and waiting for their turnaround is not, and that is the part that slips.
Skit at 80, tied with day-in-life. Same casting and direction cost, plus a risk none of the others carry: it can be executed perfectly and still not be funny. There is no version of a spec table that fails that way.

Why it is deliberately coarse

Three bands with a little spread inside them carries the real distinction without pretending to a precision nobody has. A score of 62.4 would be a lie with a decimal point on it.

And you should know what these scores don't know:

Your team. An in-house studio with a standing creator roster prices day-in-life very differently from a two-person growth team.
Your product. Some things are hard to film. A service is not a sneaker.
Your rates. Creator cost varies by an order of magnitude across categories and markets.
Performance. Nothing on this page says a hard format works better than an easy one. It doesn't, reliably. That is the whole reason cheap tests are worth running.

If your own numbers differ, use yours. The point of publishing the scale is that you can argue with it.

How to use it this week

Take your list of creative gaps, put a difficulty score next to each one, and sort by the cheapest thing you have never tested. That is your next sprint.

Score your own backlog. Go through the concepts waiting in your brief queue and mark each Low, Medium or High. Most teams find their entire backlog is High, which explains why it never moves.
Find the cheap gaps first. Cross format against angle for your category (the taxonomy page has the axes). Any empty cell whose format sits in the Low band is a test you can run before Friday.
Budget in assets, not ideas. Difficulty tells you what one asset costs. The concept calculator tells you how many you need for the outcome you want. You need both numbers to have a real conversation about creative budget.

Know what an ad costs before you brief it