HoldFast vs Loops (useloops.com) for testing video

Loops shows a respondent a whole campaign in a sequence you control, on a panel of 110 million people, priced for agencies and brands. HoldFast shows each person one version of one piece and counts only the people who finished. Different designs for different questions.

Loops is a creative testing service for agencies and brands. It puts a whole campaign, up to 30 assets, in front of a global panel and returns annotated feedback and an AI summary within hours. HoldFast audience testing shows two finished versions of one piece to two fresh groups, one version per person, and reports which one landed on a question you froze before launch. Of everything in this section Loops is the closest in spirit.

What Loops does well

Reach first. Loops is integrated with 300+ research panels, a network it describes as over 110 million people, and says you can test in any major market or language.

Then the format. A respondent sees the campaign as a whole: up to 30 assets at once, images, videos and audio files, in a step-by-step interactive flow where you control the sequence. Feedback is annotated onto the video by the respondents. Question types cover open-end, multiple choice, ranking, Likert and slider. An AI assistant summarises results, there is a Figma plugin, and it tests audio: music, sonic identities, voiceovers. The video page says 300+ people in one test in 10x less time for half the cost; the average test completes within 48 hours.

Quality is handled: a machine-learning model removes gibberish verbatim, a manual flag refunds a bad respondent, no personal data is collected, no synthetic data.

The design difference

In Loops one respondent reacts to many pieces in one sitting, in the order you set. That is the right design when the question is how a campaign hangs together.

A HoldFast test is monadic. Each participant sees exactly one version on a plain confidential grey page, watches it once with no skipping ahead and no rewatching, then answers a short survey. Only people who finished count. The verdict question is chosen before launch and frozen. Two fresh groups, two first impressions of a whole piece, nothing else in the sitting to colour the reaction. That is the right design when the question is which of two cuts is stronger.

"Backed up with statistical significance"

Loops says its results are backed up with statistical significance, and recommends a minimum of 100 respondents per test. HoldFast puts the same idea into the sentence you carry into the room. The report says one version outperformed the other only when every version reached the size committed at launch and drop-off was balanced. Below that, directional advantage. Inside what chance could produce, too close to call, which is information rather than failure.

Price, side by side

Loops publishes its prices. Starter is pay as you go at $20 per respondent on the pricing table (the FAQ on the same page says $19), 10 questions per test, 3 demographic filters, no minimum term. Team is a rolling licence: Agency from $4,000 a month with 250 respondents included, Brand from $5,600 with 350, $16 per additional respondent, 15 questions, a 3-month minimum. Enterprise is bespoke and annual. B2B respondents from about $25 each. Your own audience through a link carries no respondent fee. At the recommended minimum of 100 and the pay-as-you-go rate, a first test starts around two thousand dollars.

HoldFast: about 4 credits per person for a short recruited test, one credit being one dollar, so 24 people is around 100 credits, 48 around 200 and 200 around 800, floor 25 credits. A longer video or a narrower audience costs more per person. The exact price is on screen before launch, nothing is charged until you approve, unfilled seats are refunded at settle. A test sent to your own list through a share link costs no credits.

Who should pick which

Pick Loops when you are an agency or brand team testing whole campaigns every month, across markets and languages, on a licence that earns its keep in volume.

Pick HoldFast when you are a small video team with two finished versions of one piece and a client meeting. The same upload serves the review link, the showreel page and the test, reviewers never need an account, and the report opens with The Bottom Line, carries real quotes in In Their Words, and puts the numbers last.