I Named an Instructional Design Model After Myself — and I'd Still Change It
- Jun 8
- 4 min read
Updated: Jun 9
Week 5 · Blog Post 8 · Web 2.0-Based Learning and Performance · Summer 2026
The Embarrassing Part First
Let me get it out of the way: I built my own instructional design model, and the acronym spells my last name. DALEY — Discover, Analyze, Learning Solution, Execute, Yield. I'd love to claim this was a happy accident, five phases that just happened to line up. It wasn't. I backed into it on purpose, which puts me in the distinguished company of nearly everyone who has ever named a model (looking at you, Dick and Carey).
Here's the part that actually matters. I didn't build it to stamp my name on something. I built it because ADDIE — the model most of us learn first (Branch, 2009) — kept describing a job I don't have. ADDIE opens at Analyze, as if every project arrives with a clean question and a designer standing by to ask it. Mine almost never do. They show up as an intake form, or a manager handing me a folder of half-finished content and one sentence of context, with a quiet expectation that I'll work out what the problem even is. That isn't analysis. That's discovery.
Two Places the Standard Models Stopped Fitting
So Discover comes first, and it does the unglamorous work: where did this request come from, who owns the performance gap, and what already exists so I don't rebuild it. Only then does Analyze do its real job — the go/no-go investigation into whether this is even a training problem, or an environment, process, or motivation problem wearing a training costume. The Dick and Carey model built a whole systematic backbone around getting that question right (Dick et al., 2021), and I kept that discipline: objectives before content, assessments before materials. I just stopped pretending the question shows up pre-packaged.
The second split I'll defend to anyone. ADDIE separates Design from Develop. In thirteen years of this work, I have never once handed a finished design to a developer and watched it come out right on the first pass. Real projects shuttle back and forth — design a piece, build it, show a subject-matter expert, collect notes, redesign, rebuild. Often I am the designer and the developer, end to end. Allen and Sites (2012) made this exact case when they left ADDIE for SAM and its successive-approximation loops. I folded Design and Develop into one phase, Learning Solution, because keeping them apart is a waterfall fantasy that stopped matching reality a long time ago.
Renaming Evaluation Was the Whole Point
If DALEY has one real argument, it lives in the last letter. ADDIE ends on Evaluate, and that placement quietly teaches something false: that evaluation is a thing you do at the end. But formative evaluation starts the moment you draft your first objective. Reserving an entire final phase for "evaluation" makes it sound like a finish-line activity, when most of it has already happened by then.
So I renamed it Yield. Same slot in the pipeline, different center of gravity. Yield doesn't ask "did we evaluate?" — it asks "what did this produce?" It's the summative results check, the Kirkpatrick levels one through four (Kirkpatrick & Kirkpatrick, 2006), the return-on-investment conversation that proves the original problem actually got solved. Then it closes the loop: revise, summarize, hand off the files so the thing stays maintainable after I'm gone. Some people would still file all of that under evaluation. I don't. Evaluation is a measurement. Yield is the harvest.
Where It Holds Up, and Where I'd Cut
In practice, the structure earns its keep. Twenty-nine milestones and five approval gates sound bureaucratic until you've watched a project's scope balloon because nobody formally signed off on anything. The gates are the antidote — nothing moves to production until someone with authority says it can. The per-milestone deliverables turn fuzzy expectations into concrete ones, and, true to the moment, each milestone carries a suggested AI prompt to cut the blank-page tax.
What would I change? The weight, honestly. Twenty-nine milestones is right for an enterprise build with real stakeholder layers; it's absurd for a forty-five-minute compliance refresh due Friday. DALEY needs a documented "lite" path — same phases, same gates, collapsed for small or fast jobs — so the model bends under pressure instead of getting abandoned the first time someone's in a hurry. And I'll confess the acronym cost me something. My first draft had seven phases; I compressed it to five partly because five spelled my name. That's the tension I haven't fully resolved: a model named after its maker has to keep proving it's about the work and not the branding.
So I'll hand it to you: if you've ever bent a framework to fit how you actually work — ADDIE, SAM, or something homemade — did the change make you a sharper designer, or just a more comfortable one? I'm not convinced those are the same thing, and I'd like to know how you tell them apart.
Comments