The dawn phenomenon in my experience can be erratic as well, most days almost non existent for me but on a few days quite significant (e.g. a rise of at least 2 mmol between the small hours and waking). Having been put on Lantus, one injection before breakfast, my initial experience was good compared to Insulatard. Then, once I'd been on Lantus for some time I realised it was not covering the full 24 hours consistently, often resulting in a BG rise of 4 to 6 mmol between the small hours and getting up before breakfast, which made it difficult to optimise my dose properly. I have now split my dose 50:50 between morning and evening (7 units each, 12 hours apart) and this has sorted out the 24 hour cover problem and I can live with the ,for me at least, smaller dawn effect. Lantus' makers may claim it gives a flat 24 hour duration, but it really doesn't for some people.
The additional benefits for me of splitting the dose are a reduced tendency to become hypo just before lunch and when on long walks, being able to reduce my daily dose by 6 units and fewer correction doses of Humalog. My last HbA1c was 37 mmol/mol.
Distinguishing between different effects, such as dawn phenomenon and duration of your basal dose, can be a pain in the whatsit. You may find there are at least two variables to take into account. Best of luck and do plenty of BG testing after adjustments.
