Do you have a trip meter or are you calculating your fuel usage per tank? If per tank you may not be as aware of the little things in your driving day that increase your consumption. I have a trip meter and can watch the impact of every little thing I do, and the effect of even one off- or on-ramp on the rest of a highway trip is astonishing. If I zero the meter once I'm up to speed on the highway I can get under 8L/100 km, I think I've even gone under 7 for the odd stretch (downhill with a tailwind, most likely). But if I zero it before I get on the highway, or if I have to slow down/speed up for traffic on the highway, the average quickly climbs.
I have also tested for how bad it can get in the city: zeroed the meter at the gas station and then driven home (about 5 blocks: one red light, 5 corners, one U-turn, parking) and find it reads well over 30L/100 km. Then I go for a longer, smoother trip and it averages out. But obviously, if I include a lot of such super-short stop/go trips in a tank, the overall efficiency will be poor. But if you are assessing your driving from memory, I think we tend to sort of discount those little trips on the assumption that being short, they don't matter much. But they do.
And I have compared the per-tank readings with manual calculations so know the meter pretty accurate.
I had less impact than I expected from reaching the break-in point, but have the impression that running the air conditioner increases usage a fair bit, and consumption has been really high this past two weeks driving in snow (and in cold with lots of heater fan use, which on several settings also runs the A/C if I understand correctly).