Bonfire Labs review

137 named case studies on /work, filterable by client and year, for recognisable enterprise brands: Google, Adobe, Uber, Salesforce, GoPro, Logitech, HPE, Splunk, Toast, Ancestry, Okta, Fitbit. Work library indicates current 2024-2026 projects, not a stale reel. Individual case pages (Uber Women Rider Preference, Splunk AI Explainer) match their titles and describe the actual work. No independent reviewer commentary on craft exists to push this to Excellent.

What this dimension measures: The primary reason to hire a production agency. Look for a body of work that is current, attributable and recognisable, with named clients whose brands carry real standards. Distinctive, consistently high-quality output for demanding clients scores 4-5. Competent, generic output scores 3. A portfolio dominated by work more than two years old scores 2 — the team that made it has usually moved on. Where reviewers independently credit narrative or storytelling ability rather than just delivery, weight that: turning a complex proposition into something clear is the hard part of this job. Counts for 25% of the weighted score.

Evidence: partly checkable — corroborated in part against the sources below; the remainder rests on the agency’s own account.

No pricing, ranges, or packages published anywhere on the site; bespoke-only quoting with no public anchor. The only anchor is a $25,000+ minimum project size listed on their Clutch profile, which is self-reported to a directory, not published pricing. Scope is broad (strategy through post) with no stated boundaries of what they do not do.

What this dimension measures: Published prices, published ranges by video type, or fixed productized packages score 4-5, and an all-inclusive fee that covers script, music, narration and revisions scores at the top of that band because the quoted number is then the real number. Ranges given only on request score 2-3. Bespoke-only quoting with no public anchor scores 1-2. Score scope clarity alongside price: a vendor that states plainly what it does NOT do — no shoots, no strategy, animation only — is more useful than one implying it does everything, and should not be penalised for the boundary itself. Counts for 20% of the weighted score.

Evidence: inferred — our reading of indirect evidence, not a documented fact.

Stronger here: Epipheo scores Excellent on the same dimension.

About page describes a four-part pipeline (Strategy, Creative, Production, Post-Production) and says they customize process, talent and timeline to any project, but names no collaboration tooling, no approval gates, no revision policy. Zero independent reviews exist to corroborate or contradict on-time delivery.

What this dimension measures: Video projects fail on coordination far more often than on craft. Look for a described process, named collaboration tooling, defined approval gates, and independent reports of on-time delivery. Reviewers consistently citing clear communication and structured workflow score 4-5. No described process and no reviewer commentary scores 2-3. A vendor whose reviews mention schedule slippage or unresponsive periods scores 1-2 on verified evidence. Counts for 15% of the weighted score.

Evidence: vendor stated — the agency’s own claim, recorded as theirs rather than ours.

Both live action and animation, across many formats: portfolio service filters include Live Action, Motion Design, Brand, Event, Customer Stories, Product Marketing, Investor (a distinct IPO-roadshow line with ~10 named roadshows: Asana, Hims, Livongo, AppLovin). About page lists video/photoshoot, motion graphics, 2D and 3D animation, editorial, color grading, and generative AI. The Splunk AI Explainer combines illustration, live action and motion graphics in one piece. Video files themselves were not viewable in this text-based read, so the range is evidenced by the categorized work library rather than screened output.

What this dimension measures: What the agency can actually produce: live action, animation (2D/3D), motion graphics, studio versus location, and the format and aspect-ratio coverage needed for where the video will run. Both live action and animation, across multiple formats, scores 4-5. A single modality scores 2-3 — not a defect, but a real ceiling that the buyer must match against the brief. Animation-only vendors in particular cannot serve briefs that need people on camera, and that should be stated in the verdict rather than discovered later. Counts for 15% of the weighted score.

Evidence: partly checkable — corroborated in part against the sources below; the remainder rests on the agency’s own account.

Complete silence: nothing on the site about who owns finals, raw footage or project files, or how long music, stock and talent licences run. Must be raised before contracting.

What this dimension measures: Who owns the final file, the raw footage and the project files, and for how long the music, stock and talent licences run. This is where video differs most sharply from design: a licence that expires, or talent usage limited by term and territory, can make a video unusable in year two. Explicit ownership transfer with stated licence terms scores 4-5. Finals-only delivery scores 2. Silence scores 2 and should be raised before contracting, since retrospective clearance is expensive. Counts for 10% of the weighted score.

Evidence: inferred — our reading of indirect evidence, not a documented fact.

Stronger here: Epipheo scores Strong on the same dimension.

No stated timelines from brief to delivery and no booking-horizon information anywhere on the site. The about page's claim of customizing timeline to any project is the only mention of scheduling.

What this dimension measures: Stated timelines from brief to delivery, and how far out the agency books. Fixed, published timelines score 4-5. Multi-month cycles are normal for high-end work and score 2-3 — score the honesty of the estimate rather than the length, since a realistic twelve weeks beats an optimistic four. Undisclosed scheduling scores 2. Counts for 10% of the weighted score.

Evidence: inferred — our reading of indirect evidence, not a documented fact.

Stronger here: Epipheo scores Excellent on the same dimension.

Clutch profile read first-hand shows 0 reviews (0.0 rating) despite listing the company since founding year 1991 and 10-49 staff. Yelp page is bot-blocked and could not be read, so no rating is cited from it. No verified independent client reviews were readable anywhere. Consistent with a referral-driven shop, but a buyer has no independent reference base.

What this dimension measures: Volume and recency of independent, verified reviews. A substantial base on a verification-requiring platform scores 4-5; a handful of reviews scores 2-3; none located scores 2. Low weight deliberately — in this category a portfolio is more informative than a review count, and several excellent production shops have almost no review footprint because their work arrives by referral. Counts for 5% of the weighted score.

Evidence: verified — checked against a named source you can open; the links under Sources below are where to check it yourself.

Stronger here: Vidico scores Excellent on the same dimension.