Skip to content
Book strategy call

$ cat positioning/AGENCY-VS.md

Two models.
One ships

Most agency engagements fail the same way. Three months of discovery, a Figma file no one installs, a quarterly invoice with no commits to back it. We wrote down — column by column — what the other model looks like, so you can decide before you sign.

AWS Partner — Specializations Program
The 80/20 Hybrid™ production method · Built in Lehi, Utah · Senior engineers ex-Adobe / eBay / Venafi

# the-matrix

Side-by-side, dimension by dimension.

Sourced from a 13-competitor audit (April 2026) — Objective, Spire Digital, TELUS/WillowTree, Thoughtbot, Fueled, Instrument, MindSea, YML, and others. None published a starting price. Two required a $75k–$250k floor. We printed ours below.

DimensionAppsTangoStandard agency model
Starting price
$10,000 — printed on the homepage

In our 13-competitor audit, zero published a starting price. Two require a $75k–$250k floor.

Hidden until a sales call
Stated timeline
30-day MVP. 6-week full cycle. Locked in week 1.

Agency timelines flex with billing. Ours flex with engineering reality, not invoice cadence.

None — "depends on scope"
Team shape
100% senior engineers (ex-Adobe, eBay, Venafi)

The pyramid model is profitable for the agency. It is slow, lossy, and costly for you.

1 principal + 4 juniors — you pay for all 5
Discovery phase
1 week, ends with a fixed-price SOW

A real discovery is a Lehi day, an architecture sketch, and a number. Anything longer is billing the meter.

8–12 weeks, ends with another deck
Pricing model
Fixed-price packages from $10k

Hourly billing structurally rewards delay. Fixed-price puts the engineering risk on us, where it belongs.

Hourly, $150–$300/hr — slow is profitable
Sprint cycle
Two weeks. Working code in your repo by Friday.
Two weeks of meetings, code "in progress"
Code ownership
Your GitHub org, day one
Their repo. Transfer for a fee on exit.
AWS account
Your AWS account, your IAM root
Their account, billed back to you with margin
AI / agent strategy
Bedrock + AgentCore in production — the 80/20 Hybrid™

Most agencies are still adapting; we ship Bedrock-hosted agents to AgentCore Runtime today.

"We'll add AI in v2"
Onshore vs. offshore
100% onshore senior. Lehi, San Diego, Phoenix, Denver.

YML and similar premium shops have multi-year Clutch reviews flagging this exact pattern.

Mixed; offshore handoffs after sale
App Store submission
Included. ~95% first-try approval rate.
Out of scope. Charged separately or "you submit."
Post-launch on-call
30 days senior engineer on-call. Same person who built it.
Per-hour retainer. Ticket queue. Junior triage.
Cancel mid-engagement
End of any 2-week cycle, no fee
60–90-day notice + retainer wind-down
Cloud cost at launch
~50% lower via reusable AWS components from 100+ prior builds
Greenfield. You pay to reinvent auth, billing, observability.

# agency-anti-patterns

The five ways agency engagements die.

Each of these is a structural feature of the standard agency model — not a bad apple. We rebuilt the model so we cannot offer them, even if we wanted to.

✗ anti-pattern

The Discovery-Until-Dead Loop

What it looks like: Three months of personas, journey maps, and stakeholder interviews. A 200-page Figma file. No installable software.

Why it happens: Discovery hours bill at the same rate as engineering hours, with none of the engineering risk. The model rewards length.

✓ how we beat it

Discovery is one week and ends with a signed fixed-price SOW. Anything we genuinely do not yet know becomes a clearly named uncertainty in the SOW, not an excuse for another sprint.

✗ anti-pattern

The Junior Pyramid

What it looks like: Your salesperson met the principals. Your project gets four people you have never spoken with — three of them junior, one of them offshore.

Why it happens: Bench utilization is the agency P&L lever. Putting expensive seniors on your project loses them money.

✓ how we beat it

Senior engineers only. Same humans on the call as on the keyboard. We can do this because we deliberately stay small enough that there is no junior pyramid to feed.

✗ anti-pattern

The Hourly Bait-and-Switch

What it looks like: A "fixed-price" SOW with a scope-change clause that triggers in week three. From then on, every change is hourly, and the meter never stops.

Why it happens: The contract structure shifts engineering uncertainty onto you while preserving every billing minute for them.

✓ how we beat it

When scope grows, we write a new fixed-price addendum. We never shift to hourly billing. If the new scope is too speculative to fix-price, we stop and tell you so before signing.

✗ anti-pattern

The Offshore Hand-Off

What it looks like: The seniors you met in pitch hand the work to a Tier 3 offshore team after the contract signs. You discover this in week six when code quality drops off a cliff.

Why it happens: It is the only way to charge premium-agency rates while paying mid-market wages. The margin lives in the gap.

✓ how we beat it

No offshore. No subcontract. The engineer in your standup wrote the code in the PR. We are small, US-based, and deliberately not chasing scale.

✗ anti-pattern

The "AI in v2" Stall

What it looks like: The proposal mentions AI as a "future enhancement." Six months later there is still no agent, no Bedrock, no production model behind anything.

Why it happens: Most agencies have a single GenAI champion and the rest of the bench has not shipped a production agent. They hope AI fades.

✓ how we beat it

AI is shipped on day one when the workflow needs it — Amazon Bedrock, AWS AgentCore Runtime, Step Functions for the deterministic 80%, and Bedrock Guardrails on every output. Our 80/20 Hybrid™ is the production method, not a marketing line.

# the-shelf-thinned

The "local app dev" shelf collapsed in 18 months.

If you started shortlisting agencies in 2024, half your shortlist no longer exists in the form you remember.

  • WillowTree
    Absorbed into TELUS Digital ($1.2B acquisition)
    Floor price rose sharply. Focus blurred to enterprise CX.
  • Mutual Mobile
    Acquired by Grid Dynamics for $16.1M; site is a redirect
    Sales motion now routes through enterprise reps.
  • Savvy Apps
    Permanently closed in 2020
    Site still serves but the company is dead.
  • Kiln Lab
    "Launching Soon" splash page only
    Not a credible competitor today.

We did not engineer this consolidation. But it changes the math: the buyers who used to split their RFP across three regional senior-shops now have one. We intend to be the one that prints the floor price.

→ if you are stuck mid-engagement

The Agency Escape Plan

Free 30-minute triage on the SOW you already signed. We tell you whether to renegotiate, sue, or walk — and what the recovery path looks like.

→ the production method behind it

The 80/20 Hybrid™

The diagram, the split, and the failure modes we beat — the canonical leave-behind we use in every Lehi strategy session.

$ bring the SOW

Get a free agency-escape triage.

30-minute Lehi or Zoom call. Bring the SOW, the most recent deliverable, and the things that are bothering you. You leave with a clear next step — whether you hire us or not.

Book your Lehi strategy sessionOr scope a fixed-price quote