Integration is decided before the deal closes
The price of an acquisition gets months of attention. How the two companies will run together often gets a slide.
Every acquisition model has a line for the savings and the growth the combination is supposed to produce. Somebody will have to deliver each item on it. I have worked on more than a dozen acquisitions and divestitures, and the clearest predictor of how an integration went was whether each of those items had a person's name next to it before signing.
Decide what kind of integration it is
There are three honest answers. The acquired company can be absorbed completely, left to run on its own, or connected at a few chosen points such as finance, purchasing, and sales. Each can be right. What costs money is not choosing. A company that is independent in principle gets integrated piecemeal anyway, one request at a time from every function in the parent, and ends up carrying the costs of both approaches.
People and customers first, systems later
In the first weeks, people in the acquired company want three answers: whether they have a job, who they report to, and what changes in their pay and benefits. Give those on the first day, even when the answer is a date by which they will know. Their customers want to hear that nothing changes for them, from someone they already deal with.
Systems can wait, and usually should. Moving a company onto new software in its first quarter under new ownership stacks a second disruption on top of the first, at the moment its best people are deciding whether to stay.
One owner
When each function integrates its own piece, the work gets done function by function and nobody watches the whole. Name one person who is accountable across all of it, full time if the deal is large, with the authority to resolve disagreements between functions. That person should have been in the diligence, so the reasoning behind the price is not lost when the deal team moves on.
Assume they know something
The acquirer tends to assume its own way of working is the better one. Yet the company was usually bought because it does something well. Ask its people early what they would keep, adopt at least one of their practices in the parent, and say publicly where it came from. It costs little, and it tells the people who built the acquired business that the buyer was paying attention.
Turn the deal model into the operating plan
Most deals are judged at closing and not looked at again. The assumptions that justified the price should become line items in the plan, each with its owner, reviewed monthly for at least a year. A review at the twelve-month mark, setting what was expected against what happened, is uncomfortable and worth doing. It is the only way the next deal gets priced by people who know how the last one turned out.