Let's assume this flue gas composition in %vol:

{O2 = 5.17%, N2 = 64.18%, CO2 = 28.75%, H2O = 1.85%, SO2 = 0.05%}

For this composition, I get the average specific heat between 240°C and 400°C:

Cp(400°C,240°C) = 0.265 kJ/kg/°C

Compared to your assumption that Cp=0.24, this is a factor  0.265/0.24 = 110%  higher.

Similarly, I don't understand what you mean by "flows are correct".
If these are measurements, there must be uncertainty, it can easily be 5%.
If these are not measurement, then how are they calculated, is this based on a model, is this model assuling Cp=0.24, ...  ?

The errors you mention are not surprisingly high.
Your excel file contains 15 efficiency calculations.
I don't know their exact meaning and if their can be all treated in the same way.
Nevertheless, the average of these efficiencies is 99.3% .
Their standard deviation is 14.1% .
There is obviously a random source of error, while the efficiency is quite close to 100% in the average. The most likely source of error is the mass flow measurement. Performing more precise calculations for specific hat and density, taking gas composition into account, can only help.