The $10 Billion Question: Why 70% of DAM Projects Fail—And Why Asking “Why?” Reveals a Messier Truth

Ask people in the digital asset management industry why DAM projects fail, and most will hand you the same number: 70%. Follow that number back to where it actually comes from, and it turns out nobody has ever measured DAM failure at all. The 70% is borrowed — from general digital-transformation research whose own footing is shakier than the people repeating it seem to know.

---

Where the 70% Number Actually Comes From

The figure traces to Boston Consulting Group’s 2020 report on digital transformation, which found that 70% of digital transformations fall short of their stated objectives. That much is real. What gets dropped when people quote it is BCG’s own breakdown of that 70%: roughly 30% of transformations fully succeeded, another 44% created some value despite missing their targets, and 26% delivered minimal value. Most projects in BCG’s sample produced something. “Fell short of objectives” is doing a lot of work in that headline number, and it isn’t the same claim as “failed.”

McKinsey gets credited with the 70% figure almost as often as BCG, but McKinsey’s own 2018 global survey on digital transformations reported something starker: only 16% of respondents said their transformation had successfully improved performance and equipped the organization to sustain that improvement long-term. That’s a grimmer number than 70%, and it’s specific to digital initiatives — yet it’s cited far less, probably because “70%” had already become the standard line by the time that survey came out. The vaguer, more widely repeated “70% of change fails” claim actually predates the digital-transformation conversation: it traces to a 2015 McKinsey article on change management with no cited study behind it, describing organizational change programs broadly, not digital projects specifically.

None of this is a DAM-specific study. Nobody has run a rigorous, methodologically transparent survey asking “how many DAM implementations meet their stated goals, and by what definition of goal.” The 70% gets pasted onto DAM because DAM sits inside the broader category of enterprise software rollouts, and it sounds authoritative. It isn’t wrong to say DAM projects struggle — there’s plenty of evidence for that, covered below. It’s wrong to cite a specific, unattributed percentage as if someone counted.

---

What the Market Data Actually Shows

If failure rates are murky, spend patterns aren’t. Mordor Intelligence puts the DAM market at roughly $7.5 billion in 2026, growing toward $14.4 billion by 2031 at a 13.94% compound annual rate. Buried in that same report is a more telling detail: the services segment — implementation, migration, training, ongoing governance consulting — is growing at 14.33% a year, faster than the market overall, while software licensing (still the majority of spend at roughly two-thirds of the market) grows more slowly. People are paying more, proportionally, to get DAM working than to buy it in the first place. That’s indirect but concrete evidence that “buy the software” is not where the real work happens.

MediaValet’s 2026 DAM Trends Report backs this up from the usage side. Adoption, not feature depth, is what the report calls the strongest predictor of DAM ROI — and adoption tracks closely with how much friction the system removes from someone’s actual workday. The report found 43% of organizations saw adoption increase specifically after adding mobile access. Not after a training push. Not after a new UI. After people could reach the DAM from a phone during a shoot or a meeting instead of waiting for a desktop.

---

The Three Places DAM Adoption Actually Breaks

Search that doesn’t search

This complaint is old and it hasn’t gone away. In a long-running discussion thread on DAM News, a DAM practitioner named Spencer Harris described exactly the pattern that still shows up in vendor pitch decks a decade later: “I have watched our users type in a partial file name or project name to search for what they are looking for, knowing ahead of time there is less than 20 options out there, but yet the search results are in the thousands and in some cases tens of thousands.” The underlying cause is almost always the same — DAM search tools are built to cast a wide net across every metadata field, which is exactly wrong when the user already knows roughly what they want and just needs it fast. If grabbing an asset takes longer through the DAM than through a Slack DM or a shared folder, people will use the DM, every time, regardless of how good the system is on paper.

The integration gap

ImageKit’s research found that 100% of surveyed enterprise-scale teams (400+ employees) expected either native integrations with their core tools or a flexible, headless API — no exceptions. When that expectation goes unmet, teams don’t stop using their other systems; they build a workaround and the workaround becomes the real process. MediaValet‘s report quantifies the gap on one common pairing: 84% of DAM users also rely on project management tools, but only 21% report those two systems as fully integrated. Where integration does happen, the report ties it to a 70% jump in cross-functional collaboration and cites improved asset accessibility as the single biggest reported benefit, at 76%. The 63-point gap between “we use both tools” and “the tools actually talk to each other” is where two parallel systems of record quietly form — one in the DAM, one in whatever spreadsheet or drive people used before it.

Governance nobody owns

This is distinct from low usage. A DAM can have decent login numbers and still be quietly decaying — tags applied inconsistently, duplicate uploads never merged, access permissions nobody’s reviewed since launch. The market data above is circumstantial evidence for how common this is: if governance and cleanup were a one-time setup cost, services revenue wouldn’t keep outpacing software revenue years into a deployment. Someone has to own metadata standards, enforce naming conventions, and periodically go back through the library — and in most organizations, that job either doesn’t exist or gets assigned as a fifth of someone’s already-full role.

---

What Successful Deployments Actually Look Like

Vendor case studies are self-selected — companies that had a bad time rarely end up on a customer page — so treat the following as evidence of what’s achievable under good conditions, not a typical outcome. With that caveat:

• Viator, TripAdvisor’s experiences arm, reports saving more than 5,000 hours a year that used to go into manually locating, resizing, and reformatting assets for different markets and channels.
• Les Mills, the fitness content company, runs its DAM across 24 regions and 12 languages for 15,000+ users, and says it cut time-to-market for new content releases from months to days — without adding headcount to support its partner network.
• Simplot, the food and agriculture company, reports $2.3 million in cost-efficiency savings from automating distribution of its assets across more than 1,300 sites and channels — a number the company says is validated annually in its own internal DAM ROI report, not just a launch-day estimate.
• Forrester’s 2022 Total Economic Impact study of Brandfolder — commissioned by Brandfolder, but based on interviews with five actual customers across hospitality, financial services, employee benefits, retail, and personal care — modeled a composite organization achieving 273% ROI and $1.13 million in benefits over three years, with payback inside six months. The underlying drivers: roughly 90% less time spent searching for assets, a 40% productivity gain on creative work, and over $220,000 saved by retiring redundant legacy tools.

What these cases have in common isn’t the vendor. It’s that each one solved a specific, named bottleneck — reformatting assets by hand, coordinating 24 regional teams, distributing to 1,300+ destinations — rather than a vague goal of “getting organized.”

---

AI’s Actual Role, Not the Pitch

Every DAM vendor is currently selling AI search, AI tagging, and AI agents as the fix for adoption. Bynder’s 2026 State of DAM report found 97% of organizations say AI has already affected their content operations, and 42% describe some portion of their content as “AI-touched” — tagged, adapted, or processed by AI in some way. 98% report at least some measurable impact, mostly in the form of time saved, though only 30% expect AI to drive top-line revenue growth in the next year. It’s worth noting these numbers come from Bynder‘s own commissioned survey of its customer base, not an independent study — real data, but self-reported by people already using AI-forward DAM tools.

The clearest example of AI actually moving adoption is Bynder’s own AI Search feature, launched in 2024. It reached 1,000 customers within its first two years, having processed more than 100 million assets and over 1.3 million searches. Growth was fast from the start — more than 350 customers adopted it within the first nine months, including named accounts like Siemens Healthineers, Oatly, Vodafone, and Campari by the end of its first year. The reason it caught on where plenty of other AI-DAM pitches haven’t: it attacked the search problem described earlier — finding one asset among thousands — without requiring anyone to go back and fix years of inconsistent tagging first. Visual search on the raw image, not on metadata someone forgot to enter, is what changed the math.

What AI hasn’t shown any evidence of fixing: who owns metadata governance, whether your PIM and Shopify backend actually talk to your DAM, or whether a “differentiate” culture will tolerate a platform built for standardization. Those are organizational decisions, and no model release changes who’s accountable for making them.

---

Questions Worth Answering Before You Buy

Most DAM shortlists get built around feature comparisons — search quality, integration count, price per seat. Those matter, but they answer the wrong question first. Before comparing platforms, get specific answers to these:

1. What is actually broken today, in hours or dollars? “Our assets feel disorganized” is a feeling, not a business case. “Designers spend four hours a week hunting for the current logo file” is something a DAM can be measured against.
2. Who owns governance after go-live, by name, with time budgeted for it? Not a committee. A person, with a defined percentage of their week set aside for metadata cleanup, access reviews, and enforcing naming conventions. If that person doesn’t exist yet, the project isn’t ready to launch, whatever the sales timeline says.
3. Which specific systems does this need to talk to? Not “integrations” generically — name the PIM, the CMS, the CRM, Shopify, Slack, whatever your team actually runs work through. Ask whether the connection is native, or whether it requires a developer to build and maintain a sync job. That answer determines your real total cost more than the license price does.
4. What does success look like at 90 days, 6 months, and 18 months — for which specific team? Company-wide adoption is rarely the honest bar. A finance team using a DAM for compliance and a creative team using it for one collaboration workflow can both be successes at very different adoption percentages, as long as each number was set in advance instead of graded after the fact.
5. How much planning happens before any contract is signed? Decisions about metadata structure, ownership, and workflow are far cheaper to get wrong on a whiteboard than inside a live system with thousands of assets already uploaded.

---

The Honest Bottom Line

Nobody has produced a rigorous, DAM-specific failure rate. The 70% figure everyone cites is a loan from general digital-transformation research, and even in its home context the number is softer than the headline suggests — most of BCG’s “failures” still created some value, and McKinsey’s own harder number (16% fully successful) tells a worse story than 70% ever did, yet gets quoted far less.

What’s well documented, across separate and independent sources, is why specific deployments stall: search built for browsing instead of finding, integration gaps that spawn a shadow system, and governance nobody was assigned to own. Those three problems show up again and again, regardless of which precise percentage you attach to them. They’re also, unlike a borrowed statistic, things a specific organization can actually go check for itself — before signing a contract, not eight months after.

---

Sources

• Mordor Intelligence, Digital Asset Management Market report
• MediaValet, 2026 DAM Trends Report
• ImageKit, Digital Asset Management Trends
• Bynder, 2026 State of DAM Report
• Bynder press release, AI Search Experience 1,000-customer milestone
• Bynder customer story: Viator, Les Mills, and Simplot
• Forrester Total Economic Impact of Brandfolder (2022, commissioned study)
• DAM News, discussion on DAM search and adoption
• BCG’s 2020 report “Flipping the Odds of Digital Transformation Success” and McKinsey’s 2018 global digital transformation survey and 2015 “Changing change management” article, as analyzed and cross-referenced in Reliamag’s review of the 70% digital-transformation-failure claim

Case studies and statistics cited above are drawn from publicly published vendor materials, a commissioned Forrester study, and independent market research. Where a figure originates from a vendor’s own customer story or commissioned survey, that’s noted in the text — read those numbers as evidence of what’s achievable, not as a representative average.