This simple question has a complex answer. I am prompted to think about it a little by my preaching in a church yesterday with three services, each of differing time lengths. While I think my third delivery of the sermon was a little longer than than the first two, overall my sermon was pretty average in delivery time, and on these three occasions didn't contribute to any undue lengthening of the services (all of which, as far as I could tell, actually kept to their planned, and customary length).
(1) There is an unstated lower limit to length! Let's face it, most people going to a mid-morning Sunday service would be puzzled if not grizzly if they were out of the church within half-an-hour. While we mostly (in my experience) talk about services being too long, it is possible for a service to be too short.
(2) When people plan services there is often an associated plan re the time (whether stated or not). If the service then takes too short or too long a time, review of the service, in the light of the plan, can lead to improved planning.
(3) Context is a significant constraint on time. If, say, there are three services on a Sunday morning, 8 am, 9 am, and 10.30 am, and some time gap between services is sought, then the first two services are likely to be constrained to 45 minutes and 60 minutes respectively. Sunday lunch is still a desirable meal, especially for hungry children and teenagers, so the constraint on the 10.30 am service will be different. What is a reasonable time for a family to get away from church in time for lunch? The above sentences presume a single centre parish. There would be other constraints in a parish where each of three services is in a different location.
(4) Personal comfort is a constraint. For what time span for a service can we reasonably expect people to engage? Expectation will vary with age and stage. Mid-week services, for example, often with 95%+ participation by people aged over 75 years, are rightly among the shortest services we hold. Young students, enjoying lots of music and a mentally-demanding Bible exposition may be very happy to come to an evening service that lasts one and a half to two hours. I suggest we do well to think about what would be comfortable for a newcomer, as well as the comfort level for regular worshippers. In some circumstances I have noticed congregational numbers dropping when service length has not been constrained.
(5) Consistency is a factor. Again, speaking from experience, and taking a hypothetical "8 am, 9 am and 10.30 am" set of services, consistency helps in several ways. Consistently tightly held deadlines for the first two services is a great help to those setting up and assisting in leading the following service. A third service in the morning is not under quite the same constraint: it could vary a little in length from week to week, but normally people appreciate knowing that they will be able to choose to leave the building between (say) 11.40 and 11.45 am.
But all this involves another set of questions: how long should the individual parts of each service take ...!