Ema Gauge · The cost of not interoperating

Estimate what you spend each year maintaining point-to-point integrations and reconciling data by hand, and compare it with a FHIR layer architecture. Every assumption is yours and editable.

Your operation today

Model assumptions

Parameters are entered in USD

Cost of building one integration (USD)
Annual maintenance per integration (% of build cost)
%
FHIR layer implementation (USD)
Layer infrastructure (USD/year)
Manual reconciliation removed by the layer (%)
%

The cost you do not see

What it costs to not be able to USE the data: ETL pipelines, preparing information before it can be analysed, and the tax every piece of development pays. Softer assumptions than the ones above, which is why they sit apart.

Annual upkeep per pipeline (% of an integration cost)
%
Extra days per feature (integration tax)
Reduction of these costs with the layer (%)
%

Status quo — point to point

62,100 USD 103,000 USD

over 3 years · 20,700 USD 34,200 USD per year

Integration maintenance1,880 USD 5,000 USD
Manual reconciliation8,400 USD
Integrations implied by growth10,400 USD 20,800 USD

With a FHIR layer

28,800 USD 57,600 USD

over 3 years · 6,270 USD 12,500 USD per year

Implementation (one-off)10,000 USD 20,000 USD
Layer infrastructure2,500 USD 5,000 USD
Manual reconciliation left2,520 USD 5,040 USD
New hub connections1,250 USD 2,500 USD

Direct difference (integration and reconciliation)

4,460 USD 73,900 USD

Break-even

30 months

Computed on the conservative scenario: most expensive implementation, smallest saving.

Topology

6

possible point-to-point pairs

4

connections in a hub topology

21 possible pairs at the end of the horizon

Second-order costs

11,100 USD 18,600 USD

per year

ETL and pipelines1,000 USD 3,500 USD
Data preparation6,720 USD
Integration tax on delivery3,360 USD 8,400 USD
With a FHIR layer7,450 USD 7,760 USD
Additional saving over 3 years10,900 USD 32,600 USD

Total difference (direct + second-order)

15,400 USD 106,000 USD

The figure above is the one that holds up in a negotiation; this is the ceiling if the data is actually put to use.

See the ETL disappear: SQL on FHIR in Ema Dash →

This is a parametric estimate, not a quote. Its most fragile assumption is the reduction in manual work: if your team does not reassign those hours, the saving never materialises. Ranges are conservative–optimistic, and no single figure from this model should be treated as a commitment.

Want this analysis as a PDF?

Optional. You already have the full result; the email is just so you keep it in writing.