If I were making this decision, and none of the players had read the novel, then I would consider both the issue of railroading, and the issue of spoiling the book for them. If it were me, I would prefer to spark their interest in reading it. That puts me on the side of creating a path for the characters that is unrelated to the story told in the book(s).
When in a situation where I will be running settings tied to films, or books, or that have invasive metaplots (Star Wars, Neverwhere, Serenity, oWoD, etc), I have two approaches:
Approach 1: Normal Workload (Just using the setting)
I try to run things the same way I would run a game without a strong primary story: that is, I would put the focus on the story of the PCs as if it were the primary tale. The story from the novel or film is placed in the background as extra detail. The events in the game may or may not ever involve the events or characters of the original source material - and I would work to keep them separate if at all possible.
The goal of this approach is to create unique adventures in an established setting that all or part of the group enjoys, without worrying about existing story lines. This can be done by simply ignoring/deleting the original story, by avoiding it by setting your stories in another time or place within the setting, or by creating unrelated plot threads with no crossover with the original story and making them the focus of the campaign.
Approach 2: High Workload (Creating subtle links to the Original Story)
If I have the time, or motivation from the players to allow their characters to get involved in the primary plots from the original source, I like to remember that historical records and memory are very selective. That perspective is the tool that I use to create subplots and important interactions with the main events, which do not threaten or change the continuity of those events.
This sort of thing involves approaches such as providing roles for the PCs which make the tasks of the characters from the original source material possible. Without the PCs' actions, the course of events in the source become impossible.
example: An information gathering quest for beginning characters
An advisor to a character from the books reveals information about the enemy, a prophecy, etc in the original storyline. As I am looking for neglected or glossed over parts of the story to flesh out, this sort of scene might catch my eye as an effective way to introduce starting characters, and enable the players (who may not have any familiarity with the sources) to establish an understanding of the setting and culture being portrayed in the campaign. To make use of it, I would:
- Create a plot where that advisor had recruited one of my NPCs to uncover information in an earlier time period.
- That NPC would then select the PCs and send them on a dangerous journey to collect that information
- Use this to establish the skills and personalities of the characters, connect them in a very tangential way to a peripheral character in the original source, while at the same time making them a vital, but invisible part of the original tale
Doing this for the course of the series, will take a great deal of familiarity with the original work, and a lot of planning to find ways to have the characters support the main plots of the source material without ever being 'deemed by history' to be a major character in the story.
The goal of this approach is that when the players read the books, they can get an extra smile and surprise when they realize their hidden role in the tale that is unfolding before them.
I am the one who brought Sir Fancy-Pants the news that the Dragons had awakened! I almost died that day...
Actually, I happen to be in a similar situation as you. My group has played games that vary in pace from Shadowrun games where we spend three real time weeks (one session per week) planning for a single run, to Pathfinder games where we clear a new dungeon every session.
Right now, we're playing a D&D 5e game set in a Dying Earth styled setting in which the world is probably going to end soon. We're all very close friends, so I think I can accurately represent some techniques the GM is doing:
Mention a prophecy. It doesn't have to have many details, or even be completely accurate, but giving the players some idea of what the end of the world is going to actually look like helps. That way, they'll know the world isn't about to end yet since they haven't seen the events of that prophecy unfold. Include several different omens, although not every omen has to be apparent right from the get-go. You might show an unexplained omen, and later have an NPC explain it to the PCs if you want the opposite effect ("oh crap the world is ending sooner than we thought").
Have NPCs prepare in a meaningful way. The players can use NPCs more well-versed in the aforementioned prophecy to gauge how close the end of the world is. In our particular example, the queen of a city-state has decreed that the citizens are to build a citadel on a particular hilltop to prepare for the end, since that's mentioned in the prophecy. The players can get a very rough idea of how close the world is to ending by knowing whether the citadel has been constructed or not; that is, the fact that an expert on the prophecy hasn't hidden away yet should mean it's still safe.
You can also take a more direct approach. If you notice that the players are getting too anxious, you can introduce a sagely NPC that feeds them more information about the prophecy, proving to them that it's not going to happen yet. In our case, we found a hermit in a myserious lighthouse in the desert, filled with clockwerk mechanisms. Turns out that this hermit has lived for perhaps hundreds of years, maintaining the beacon, and it also turns out that he's looking for a successor, because the beacon drives away aliens bent on destroying our civilization. They believe that as long as the beacon is lit, an ancient empire capable of defeating them still stands. So, in this case, the players can reasonably be assured that they're safe as long as nothing happens to the beacon.
Best Answer
Work WITH your players
If I'm going to sit down and play a game for several hours, nevermind for several sessions of several hours, I want to know what genre I'm playing. The "gotcha" campaign doesn't make people as happy as folks think.
Tell your players, "I'm planning on doing a zombie outbreak game, and the expectations are that your characters are going to be built like (Normal civilians/civilians with some survival skills/police military etc.) and the first part will be before the outbreak."
What's to stop players from doing stuff like building a bunker before play? Because you, as a group, agree that's not the kind of zombie game you want to play, and if they don't want that, they're not going to like this campaign.
Easy.
Motivation & Relationship Mechanics
This is one of those places having relationship and ideals based rewards would work great.
When a media wants to show you more of the pre-outbreak, they show the protagonist's normal lives - the people they care about, their ideals, hopes and dreams and struggles. In games with these kinds of reward mechanics, you spend these scenes loading up these relationships and ideals, getting either XP or some kind of hero points in the process, and then when all hell breaks loose, then the players get more points for trying to protect, rescue, or reconnect with those people, places, and things.
These kinds of mechanics work very well because they set up a long term survival goal into the rules - in the short term, doing what's immediately best for survival is good, but in the long term, you want those XP/hero points to stay alive as the situation worsens.
Which...mirrors a lot of zombie fiction - the last survivor(s) end up having to balance caring for others and being actually a decent human being vs. pragmatic "I'm sorry, you've been bitten, I can't let you in.".
Social Mechanics
Games with good systems for social mechanics would also be good. Zombie fiction is a lot about the arguments about what to do next and everyone getting at each others' throats and sometimes people making bad decisions. Good social mechanics gives people a chance to influence each other, to convince each other to help, or to cut bad deals.
Tying it together
When you have the two above ideas working together, you can use the pre-outbreak time to set up relationships, rivalries, tensions and promises. All of that becomes critical when the disaster strikes.
You'll notice all that kind of stuff is what makes the best zombie stories work. It's about people falling apart and falling together that makes it work, not "I am tactically sound and have food supplies" kind of stories.