|This article is about the current version of DF.|
- This is not a tutorial, a FAQ, or a new players guide. This is a mildly advanced theoretical treatment for someone ready to take the plunge and make all the decisions about their own fortress. The following are intended for beginners:
- For an explanation of the interface for starting out, see Embark.
A starting build is a personal strategy for choosing the initial supplies, equipment, and skills of your initial seven dwarves when starting a new game in fortress mode (see Sample Starting Builds). These skills and items which you assign to your dwarves will have a large impact on life in your new fortress, especially in its first year.
This page attempts to provide a discussion about how and why you make the choices on what you bring with you. This page is not an explanation of the mechanics of doing so, see the embark page for an explanation of the interface itself. This page assumes you have already made certain decisions, such as where you plan on settling, and that you are looking at the Prepare Carefully screen.
One thing should be made clear - there is no "best" build, no "perfect" or "clearly superior" final mix of skills and items, if only because there isn't any one goal of play. The goals you have for a fortress will dictate which sets of items and skills are best suited to achieving that goal - in your opinion. And then there is the environment, where your dwarves will arrive, the creatures, the resources available, and so forth. Finally, some people do things solely because it is hard, and that makes it more fun for them.
 Components of a Starting Build
There are two components of a starting build: skills and items (which includes animals).
Skills for your initial dwarves determine how quickly they will work early on, what industries you are able to start with skilled workers, how well you can defend your dwarves early on, and what the quality of various goods they produce will be, and possibly many other considerations. This page considers in detail how you might go about choosing skills for your starting dwarves, and examines the multiple competing perspectives from which you can make that decision.
The items that your dwarves bring with you can be tailored a number of ways. There are trade-offs to cheaper and more expensive alternatives, and reasons why you might choose either. This section explores the nature of these trade-offs and the reasons for making a decision. It also looks at optimizing goods brought in more general contexts.
Starting builds can and should vary based upon a number of other variables. Where you choose to settle will dictate what raw materials are available, and thus require different skill sets to utilize, not to mention different threats from native wildlife based on biome, savagery, and alignment. Which dwarven civilization you chose will restrict the materials with which you can start. Making choices about these variables is not part of a starting build. What you choose for skills and items because of these choices is part of your starting build, and so some general guidelines about different environments is given last.
This page does not cover the interface for accomplishing these tasks; please see the embark page for that information.
 Skill Optimization
With only 7 dwarves, you can't take every skill, so you must balance what you do take. At the starting phase, each dwarf can only be assigned a maximum total of 10 skill levels, with no single skill starting higher than "5". Maximum skill distribution is thus constrained to 1 level in each of 10 skills, or 5 levels in each of two skills, or something in between. Because dwarves can learn any and all skills once your fortress starts, these initial choices do not dictate what the dwarves can do, opening up incredible latitude to choose skills for reasons other than survival.
- (* Note that an unskilled dwarf starts with all Skills at Level 0. Adding +5 Levels is then Level 5. This is true regardless of how many "points" a level costs when first buying skills at embark.)
A brief list of considerations governing skill choice:
- Maximizing starting skill ranks vs. generalizing and having more skills covered at lower levels
- Balancing multiple skills for a single dwarf, so they aren't constantly needed for two different tasks at critical periods
- Military vs economic needs
- Your goals vs "basic survival needs" to keep your fortress healthy and happy
- Speed that a skill can be trained in game
- Demand for a skill during a game
- Whether quality or speed are significant considerations for tasks/final product
- Balancing the desire to create wealth (with high-value products) with the need to maintain morale (with low-value but commonly used products, like beds, which normally are made from wood)
- most importantly - your playstyle - what you think is "fun"!
 Breadth vs. Depth
A dwarf with only 2 starting skills at 5 ranks each is pretty good at 2 tasks, but untrained at anything else. A dwarf with nothing higher than level 1 is passable at many tasks, but not good at anything. Each dwarf in your party is going to be somewhere along this continuum, and you'll need to choose where.
Level in a skill dictates how fast a dwarf completes a task (most of the time), and the quality of the result (if applicable). On one extreme, butchery has no time variance for slaughtering a tame animal, and has no quality associated with the outcome. On the other extreme, metalworking tasks can take a long time for an untrained worker to complete and their high material value means the quality multiplier has a large impact on the end value.
Dwarves who complete tasks faster can do more total jobs within a given time. The rate at which speed increases with level varies with skill, so some skills will benefit more than others.
Dwarves who make items of higher quality will contribute more to fortress wealth and may have a large impact on fortress happiness if their work is readily available to be seen by other dwarves. Items which typically contribute to happiness are common public items like beds and tables. Items which contribute the most to fortress wealth often cannot be displayed, but make useful trade goods or equipment for your military.
Aside from the obvious trade off, there is another reason to prefer depth - dwarves can obviously only complete so many total jobs within a given timespan. If a dwarf is busy doing one thing, he can't simultaneously be doing something else. So a dwarf who is highly skilled in a few skills may not actually experience any disadvantage if he is kept doing those things in which he specializes. The generalist dwarf, on the other hand, may be able to do many more different tasks adequately, but he can still only do one type of task at a time. A dwarf with one highly used skill (such as Mechanics or Mining) can feasibly spend all his time using only his primary skill and thus has no need to generalize. In effect, the generalist is wasting more skill points whenever he does jobs than the specialist, so long as the specialist tends to do jobs he has levels in. Specializing your initial skill investment is therefore superior if you specialize the division of labor in your fortress.
Of course, you can still only bring 7 dwarves with 10 total levels of skills each, so covering everything you want to do in 14 skills may be hard, if not impossible. A generalist or two can cover more bases that have little quality need or are otherwise fast even without a high level. The generalist's real problem arises from the fact that any dwarf can do any task, and having 1 level isn't much better than having no levels. Which isn't to say there isn't a situation where a 1/1/1/1/1/5 dwarf is the right solution (indeed, the typical recommended leader/broker takes 1/1/1/1/1 in appraiser/judge of intent/negotiator/+2 social skills because none of these skills have a time or quality component), but most less-specialized dwarves are more likely to fall in the 5/3/2 or 4/3/3 end of the spectrum solely because there is a minimum investment necessary to be noticeably better than not having any levels at all.
 Design Constraints: Which skills do I need, really?
The only thing that you absolutely must do in the first year is get your food supplies into a food stockpile, preferably inside, or your food will rot on the ground and your dwarves will starve. Anything else you want to do can be accommodated by sufficient investment in initial supplies and/or skills. This means the options for possible starting builds are vast because virtually any set of starting skills for your dwarves is viable (and that's before you even think about equipment, which adds more variables). So the short answer is: none.
That said, there are some skills which will be used, to one extent or another, by virtually every fortress - but that doesn't mean you need or even want to invest points in them to start. You could even manipulate the fortress (see challenge) to completely avoid one or more of the following, but these are the skills you will find it exceptionally hard to avoid creating jobs for:
- Mining - to dig your fortress, and create loose stone for projects. It's only possible to avoid mining if you're secretly an elf.
- Carpentry, Wood cutting - beds can only be produced from wood (rare moods aside). This skill is also quite useful for producing bins and barrels.
- It's best not to assign carpentery and wood cutting to the same dwarf. Doing so creates a bottleneck early on, when carpentry is your only reliable industry.
- Masonry - to fashion dwarven furniture from stone, and build buildings and constructions from stone.
- Due to the vast supply of loose stones, masons can be easily trained by repeating jobs at a mason's workshop.
- Growing - your farmers' work echoes throughout so many other tasks, it's stunning.
- Skilled planters produce larger stacks of crops, which means more food, more booze, more cloth, more dye, ....
- While it's possible to feed your fortress on nothing but caravan goods, you'll never come by enough alcohol that way, so you'll eventually need to grow or gather plants for brewing. Dwarves will literally go crazy if forced to drink nothing but water for long periods. Thus you'll want to plan for farming eventually - not that you must bring a highly skilled Grower, but it'll certainly be very helpful.
- Planters with low skill levels can decrease your seed stockpiles - each seed planted by an untrained grower will usually result in a single replacement seed, but may produce no seed at all if the crop fails. When you start out with only a handful of seeds, a few losses can easily impact the sustainability of your fortress.
- Brewing, Cooking - A skilled brewer produces alcohol more quickly. Likewise, a skilled cook prepares meals more quickly (and more appealingly). However, most food can be eaten raw, and your dwarves can exist for a time on water.
- Mechanic - If you want traps, and most people will. Also needed for most machinery. Mechanisms sell for a high price as a bonus.
- Building Designer - Mandatory for some buildings and constructions, but skill only improves speed a tad and increases structure value.
- Appraiser - you will use this whenever you trade with a caravan. Without it, the game won't tell you how much anything is worth, making trading difficult. It's almost always recommended to start with a dwarf with Novice (1 pt) Appraiser skill.
- Clothier, Leatherworker - At least one of these skills will be necessary to supply your dwarves with new clothing. Articles of clothing count as finished goods with quality modifiers, and clothes can be traded easily (either before or after your dwarves wear them).
The very fact that you will use these skills can make many of them desirable to choose as starting skills for your dwarves. Of the above, Mining, Masonry, Growing, Cooking, and Mechanics are generally worth considering as "highly desirable". However, any skill can be used untrained, and/or get trained on the job - it just means a slower process and/or average lower quality product than if done by a dwarf with a higher skill level. Some skills (e.g. Record keeper) are rarely worth investing initial points in even though you will almost always use them; you can simply let the dwarf learn on the job. Others (e.g. Mining, Carpentry) may be worth investing points in depending on your goals or the tempo with which you want to achieve them. (Mining is easy to train so you could forgo initial investment and just train on the job, but this could force you to spend longer on the surface, increasing your exposure to dangers. A skilled carpenter can contribute a lot to fortress mood, but won't produce much value; see the discussion of Quality below.)
Many other skills are optional for your fortress; you will have to deliberately choose to use them. The skills above are essential to basic aspects of the game, and avoiding one requires a deliberate choice not to use it (and likely a lot of effort spent to avoid doing so). A fortress can make its wealth by smithing fine weapons, weaving quality cloth, encrusting precious stones onto furniture, or crafting quality trinkets. Or all of those. But there usually isn't a compelling reason to choose any one in particular. A fortress that never designates a tile for mining, however, requires exceptional effort to achieve.
There is no universal design constraint on which skills to start with. Ultimately the answer to "What skills do I need?" is "Whichever you want". Choosing a mixture of these commonly used skills and your desired specialized skills will make starting up your fortress easier and more efficient, but you don't need to start with any of them. Choosing to avoid some skills may force you to use some others, but nothing compels you to invest in any skill in particular.
A common skill list (Just as a general quick start):
- 2 Miners
- 1 Woodcutter
- 1 Grower/Cook
- 1 Grower/Brewer
- 1 Carpenter
- 1 Mason
This is not the be all end all, of course; it all depends upon your location, your goals, and what you consider fun.
 Balancing military and economic needs
Not all embarks will require a military presence in the first year, but anyone planning an expedition to a sinister, haunted, or terrifying biome would be foolish not to be prepared for nasty dwarf-killing creatures. The solution doesn't strictly need to be military skills; quick delving and a skilled mechanic may be sufficient; but starting with a military dwarf will give you the earliest possible protection and a lot more versatility in where that protection can be applied. Whatever you choose to do, understand the risk and be prepared for it.
 Training considerations
Some skills are harder to gain experience in than others - requiring valuable resources or taking an extended period of time, and thus inconvenient to train from scratch. Investing in some of these skills for your initial dwarves can make those industries much less painful to start. For example, metal-related skills generally eat metal bars, and thus the less time you spend training metal workers up to a decent level, the faster they'll be churning out high-quality items and the fewer raw materials (bars and fuel) they'll waste in training. On the other hand, despite its importance, skills like mining train relatively quickly and barring extenuating circumstances (expected need to accomplish particular digging projects in the first month or you'll get mauled by a Giant for example) there's little need to actually invest your starting skills in it - they can learn on the job.
 Quality, value, and happiness
Quality is an important part of Dwarf Fortress. Higher-quality items produce better and more frequent happy thoughts and are worth more money in wealth and trade. When choosing skills that produce objects of quality, the desire to produce valuable goods for trade will often conflict with the desire to produce objects that will make your dwarves happy. Built items that are frequently encountered tend to be things like furniture (especially beds) which tend to have low material values and thus low total value despite the quality of the work. Further, these things tend to be inconvenient to trade (due to weight and storage constraints). It is often best to strike a balance between dwarves who produce valuable trade goods and dwarves who produce quality items that will make your population happy - and thus be able to achieve both goals simultaneously.
It's worth noting that built furniture and worn clothing counts its value twice -- once under the appropriate category and once for displayed value. If you're trying to maximize your created wealth total, a good metalsmith producing statues from high-value metals is optimal. A mason can also build furniture directly from metal ores such as gold nuggets, if these are enabled in the stocks "Stone" screen. While building with ore saves time and fuel, it generally results in a 25% reduction of total material value compared to refining the ore. This is balanced somewhat by the fact that masonry is much easier to train, and therefore more likely to yield a high quality modifier to offset the reduced material value.
 Moodable skills
Strange moods can give a dwarf Legendary skill in his/her highest-level "moodable" skill, and moods take hold of dwarves with different professions at different rates. Some skills are "moodable" where others are not. You might choose to take a certain skill solely because it opens up moods for that skill with that dwarf. Some moodable skills are more valuable than others - a legendary weaponsmith is both valuable and useful. A legendary tanner is generally a waste of a mood since tanned hides have no quality.
Because a dwarf can only have a strange mood in one skill, pairing a moodable skill with a non-moodable skill can ensure that if the dwarf has a mood it will be in the skill you desire. See the section on combining skills below for more details on ways to pair skills.
Dwarves with no moodable skill can be allowed to do one task using a moodable skill to give them a moodable skill with no starting build investment, so moodable skill considerations should not be considered a primary reason to choose particular skills - you should also want to make use of them for other reasons.
 Combining Skills
Every dwarf is going to have 2 or more skills. This means that even once you know which skills you want, you're going to need to pair them up before assigning them. Not all skill combinations are equally functional.
Some skills are highly time-consuming, either because the skill is in frequent demand (e.g. mining), because it requires extensive travel (e.g. wood cutting), or because it takes a long time to do an individual job (e.g. strand extraction). If a dwarf is spending most of their time using that skill, they aren't making much use of their other skill. Pairing two time-consuming skills together therefore tends to be a bad idea, as one or both jobs are not going to get the attention they need or deserve. Similarly, pairing a skill with time-critical jobs with a time-consuming skill also tends to be a bad idea. If your grower is also mining, he may not stop to plant crops one season. Or he might neglect to harvest your crops in a timely fashion and they could rot on the ground (if you only let your growers harvest). Arranging your skill combinations to avoid these situations is generally beneficial. For example, Masons, miners, growers, and any craft that your fortress will base their economy off of (glass, stonecrafts, armour, etc.) will take a lot of time, so plan accordingly.
Instead, pairing relatively time-intensive tasks with less time-intensive tasks will let your dwarf accomplish all such tasks adequately. Once you get a metal industry rolling, an armorer/weaponsmith will need to make a lot fewer weapons to outfit your soldiers than he will armor components. Thus he can usefully do both jobs without hurting your productivity too much. Similarly, a mason might also be your architect, since building designer is a very infrequently used skill.
Working at different jobs levels up specific attributes. One could level up a miner until he becomes mighty and ultra-tough - and then turn him into a soldier, or retire him to haul stone. If you plan on doing so, it may not be a good idea to give this guy a second critical job that will demand a lot of time away from their focus.
Since tasks will take place in specific areas, another approach is to combine tasks into dwarves who will take care of a specific industry, or spend all their time in one generally narrow part of the fortress - the forges, or the kitchens, or outdoors, for instance. So combining Farming with cooking, rather than mining, for example, and turn on only Haul Food, not Haul Stone. Metalworkers spending their time in the forge can easily handle more than one type of metalworking skill, and are also well-situated to be furnace operators.
Similarly, you can also make the craftsmen of your finished products also responsible for the production of intermediate products from raw ingredients. This way when they run out of materials to make into finished goods they can immediately switch over to working raw products into intermediate products so they'll have more to work with later. This works better in some industries than in others. A single butcher/tanner/cook trying to process multiple animals simultaneously will likely result in rotten food, carcases, or skins. But a weaponsmith who doubles as a furnace operator can usefully ensure he has material to work with when you want him to. Later on, however, a highly skilled craftsdwarf is often better suited at sitting in their workshop and having others deliver raw materials to them than going out and obtaining their own raw materials themselves, but in the early game dwarf-time is limited, and a single dwarf who can work an entire production chain can do so relatively efficiently and let your other dwarves be used elsewhere.
There is no requirement that a dwarves job combination needs to look 'right' or logical. A weaponsmith will most probably not spend nearly 100% of their time creating weapons - what they do with the other part of their time may have nothing at all to do with forges or smithing. Jobs which require little time in general, or little time early even if time-intensive later, may well be paired with any time-intensive task solely to provide the dwarf with something to usefully and skillfully do with most of their time, and freed from that duty as needed for the other tasks.
Another constraint you can impose on your skill combinations is to try to limit dwarves to moodable skill and one non-moodable skill (or a moodable and a less desired moodable skill at lower level), so any mood will improve the desired one. For example, pairing craft skills with farming skills gives you dwarves that will perform useful food production or raw good processing services while also getting their mood in a valuable finished goods skill. Example: Clothier/Grower.
Finally, when combining skills think twice about placing your most valuable skills (e.g. grower) alongside dangerous skills (e.g. hunting, wood cutting, etc.). Your most valuable dwarves should be kept safe inside your fortress; less valuable dwarves should be chosen to venture into the wilderness.
 Which dwarf should have which skill?
Dwarves who craft goods they prefer, or work materials they prefer, gain a bonus to the quality of the finished work. This can inform your choice of which skills you choose, for example by choosing a weaver because you notice a preference for sheep wool yarn, or you might choose the skills you want and then try to find a suitable dwarf to use that skill. In the latter case, since all dwarves have one metal preference you might assign an armorsmithing skill to a dwarf with a preference for iron, steel, or adamantine.
The dwarf with the most social skills will end up being the Expedition leader, who will then become the mayor and start making demands and mandates. Thus you should avoid giving the most social skills to a dwarf who has preferences for things like adamantine and traction benches. Ideally, a dwarf with no item preferences will result in a mandate-free noble.
Dwarves have physical and mental attributes that affect the performance of certain skills. You may wish to give a socially adept and patient dwarf the leadership/broker skills, or a dwarf who doesn't tire easily a skill which will be in frequent use like mining. You can also try to match skills to personality, some of which have obvious implications for their willingness to work long hours or how frequently they might take breaks.
Of course, all combined these represent a lot of possible constraints on where you assign particular skills, and it would be impossible to apply them in total to your entire desired skill load. Use these as a guide, but don't be upset if all your dwarves are anti-social psychopaths - someone still needs to be the leader, after all.
 Other considerations
Migrants can and will arrive with a wide selection of decently trained skills. While it is a gamble, chances are pretty decent that migrants will arrive with a highly trained skill that is also highly desirable and would usurp the job of one of the seven starting ones. The first few migration waves are likely to give you a much better talent pool than what you can assign at embark. On the other hand, you may never see a migrant arrive with the skill you really want. Even then, migrants still provide an excellent pool of workers to train for the specific skills you desire.
Skills atrophy if not used (they are marked "rusty" and later "very rusty"), and they can eventually decrease in level. Consider that skills which you will use years after embark are going to be rusty or even deleveled. Embark to the first caravan is long enough for a skill to start rusting, so you might want to make sure you'll use every skill you embark with before the first year ends to avoid catastrophic rusting.
Remember that you need to survive in order to accomplish any goals. Have a plan for lasting to at least the first caravan, if not one for longterm sustainability.
The items you choose to bring with you will need to satisfy a number of needs. Most importantly, you need to keep yourself alive - at least until the caravan arrives in the fall to resupply your fortress. You probably also want to plan on some way of making a shelter, whether that be the traditional delved hall, a majestic castle, a log cabin, or something even more exotic. You may want to plan for mishaps by bringing essential medical supplies, especially those which may be hard to acquire on site. And you might bring items which will assist in creating items for trade to that first caravan, should you need anything for the skills you're planning on.
For the purposes of this article, livestock are considered items.
All embarks get the following items without paying for them: 2 animals (who pulled the wagon), and the 3 wood that make up the wagon.
A single dwarf eats about 2x/season, and drinks about 4x/season. With 7 dwarves that's ~approximately~ 14 meals per season and 28 alcohol per season, or ~42 meals and ~84 alcohol until the end of Fall. The dwarf caravan tends to arrive in the third month of fall, so you will probably need to plan on a full 3 seasons. You are also likely to get at least one if not two small waves of migrants before the caravan arrives.
It is possible to bring enough food and drink to make it to the caravan - indeed, bringing enough food isn't especially hard (especially once you factor in slaughtering the animals who hauled your wagon). Bringing sufficient alcohol is harder, although bringing plump helmets to brew can significantly cut the cost.
The likely best way to keep your dwarves in drink is also the most labor-intensive - setting up farming in the first season or two is perfectly plausible, allowing to grow your own plants from seeds and brew the products. (Keep in mind not all plants can be brewed - don't plant dimple cups and expect to make alcohol). In addition to the necessary seeds, starting your own farming operation is going to require either some soil or some way to get the ground muddy.
It is possible to make all your alcohol by harvesting aboveground plants, if highly inefficient. It also only works in biomes with collectable plant life. Notably evil biomes and glaciers are unlikely to provide suitable vegetation.
Barring a convenient cave, you're going to have to do something for shelter. Shelter is your first defense against roving creatures, keeping them away from where your dwarves are working so they don't spam job cancellations and strew items all over the place. (As you might guess, most 'convenient caves' aren't actually that convenient, as they tend to have residents). Basic walls that allow you egress won't stop a dedicated invader, but you probably won't see those until year 2 or 3, so you have a little time to develop more elaborate defenses.
Food outside will also spoil a lot faster than food inside, so making a cellar of some sort to store your food in will increase the longevity of your food supplies. The rate at which food spoils depends on ambient temperature, so the urgency of making a cellar will depend on where you settled. It might be possible to go without a cellar in a freezing biome.Verify The only way you can avoid thinking about food storage in the first year is if you collect food and make alcohol as needed - i.e., by using an herbalist to collect local plants - which can avoid needing to mine at all.
Delving a shelter requires mining, which means having picks to dig with. One can always bring one or more picks at embark, but its also possible to bring the supplies necessary to make them. See finished product or do it yourself.
An aboveground shelter can be made with stone or wood or possibly more exotic materials. Stone of course requires mining, and thus picks. Wood can be had with an axe assuming trees are present, and axes, like picks, can similarly be brought at embark or made on site. It is of course possible to bring sufficient raw materials to build walls and a roof with, but this is far less efficient than just bringing a pick or an axe, although it could make a fun challenge.
Most industries require little more than materials you can collect at the site and a workshop. As long as you have access to some sort of building material (stone, wood, or even ice), you won't need to bring anything for these. However, if you want to get an industry going immediately, it does help to bring a few building materials along (or be willing to use the wood from the wagon, if only temporarily).
Some industries require fire-safe materials to build with. Nearly all stone qualifies, as does metal. Wood can be converted to a fire safe material by burning it to ashes in a wood burner's workshop, but of course that workshop requires a fire-safe material. If you're mining, this condition is easy to satisfy, but if you intend to run any of these industries right away you will need to bring appropriate materials.
Some industries require plant or animal matter to work with. Clothiers ultimately need cloth, which comes from certain plants or animals. Leatherworkers need tanned skins. (And while you can get 2 off your pack animals, this isn't sufficient to run an industry). If you plan on running these types of industries you will need to provide suitable raw materials. Hunting can cover leatherworking needs (although this requires a hunter and hunting implements), and foraging can find rope reed plants, but its usually better to bring enough appropriate animals or plant seeds to have a good shot at getting started in a predictable and sustainable way. Similarly, milking and cheese-making require milkable animals, and bonecarving requires a dependable source of bones.
Metal industries require metal and an anvil. You cannot make an anvil on-site without already having an anvil, so if you plan on doing any forging before the first caravan, you will need to bring one with you. Metal can be brought as bars or as ores to be smelted in a smelter into bars, or can be mined yourself. Each unit of ore smelted will produce 4 bars of metal, so there is definitely a cost-advantage for creating the bars on-site. You will need to provide fuel or magma to run these workshops; a bar of coke and some bituminous coal can bring your metal industries up to speed much faster than relying on charcoal. Keep in mind that if you wish to produce steel, you will need some form of refined fuel even if you have easy access to magma. Bringing along some bituminous coal is a cheap and efficient way to guarantee a supply, especially since volcanic regions typically lack coal resources. This will greatly ease your wood consumption as well.
Soap requires wood and a source of tallow to be produced. Lye can be bought at embark to skip the first steps and make soap more directly. You will still need to bring or make buckets and have an empty barrel to actually produce soap, this is a matter of having sufficient wood. Tallow can only be acquired by hunting or from livestock you bring but both sources hamper a quick industry start anyway.
Jewellers require gems. Cut gems can be brought at embark, but are too expensive to bring in quantity. Generally a jeweller requires mining to find sufficient gems or a glassworker to produce raw glass to work with.
Glassmaking and Pottery requires sand or clay, and fuel. It's hard to viably run these industries solely off imports. You can check the pre-embark screen for clay layers, but sand is reported as a soil layer which may contain no sand at all unless in a Sand Desert biome. Like metal workshops, coal can be brought to substitute for fuel fairly efficiently, and magma can abate the need for fuel entirely.
As is probably obvious, certain industries depend on similar inputs. Planning on a set of industries which require similar complementary inputs can let you more efficiently spend your starting points at embark or more efficiently plan your digging during the first year. If you plan on a lot of fuel-dependent industries, it may be worthwhile to prioritize finding a source of magma.
 Container mechanics and free items
Many items come in containers such as barrels and bags, including food, liquids, seeds, and powders. The cost to embark with these items can be cheaper than the cost of the container itself, and each different type of item for each category will come in its own container. Furthermore, you'll get a new container after every 10th instance for food, most liquids, and seeds, and after every item of powders.Verify Alcohol gets a new barrel after every 5th unit. (Food actually groups by animal type, so if you get horse tripe and horse meat they'll combine in one barrel, but horse meat and donkey meat will come in separate barrels). Thus diversifying your initial food supply with 1 of each low-cost food item will net you a large number of barrels. Similarly, it is worth taking 1 of each seed you weren't planning on taking more of, solely for the bags. Taking some sand or gypsum powder is also a cheap way to get bags. Lye (for soap) and milks can be brought for more barrels - and milk can be made into cheese for a low-cost embark option that becomes food.
As stockpiling and some jobs are container-limited, getting as many free containers as you can will free up labor (and possibly valuable materials) that would otherwise be used making containers. Note, however, that this behavior is considered by some to be an exploit since it provides substantial advantage at no cost.
 Finished product or do-it-yourself
Rather than buy an item before embark, you can buy (some of) the raw materials, and just make the item after the game starts. The trade off is time lost vs cost saved - how long will it take you to make what you need, how soon will you need it, and how much "other stuff" can you buy instead?
The most common scenario involves forging your own picks and weapons. While not usually too much of a hardship, it can be dangerous to delay having any weapons if you expect possible hostile creatures. Furthermore, you will lose time - possibly 1/4 to 1/2 the first month - if you forge your own picks. However, if your weaponsmith is skilled (even just a couple ranks), the finished item can be much higher quality than what you can buy (which is always "lowest quality"), which is important for weaponry, but makes no difference for mining. Dwarves can fight unarmed (especially small wild animals), so you have to balance the expected danger of your selected embark sight against the savings.
Another consideration is how difficult it will be to find and gather the necessary resources on your map. Bringing a few logs/stones/blocks lets you start building some workshops immediately upon arrival, with fire-safe stones/blocks you can build a furnace and smelter, and with an anvil a forge. If you want steel weapons, bringing some iron and flux stone will almost certainly be faster than mining deep for it (and definitely will be if you have an aquifer to deal with).
You can also easily plan on making all or most the booze you'll need. Plump helmets can be bought at embark and brewed at a still, as can wood for making barrels (which are also required). Buying just a few units of each booze will be enough to hold you over until you can brew-it-yourself, and at worst your dwarves can drink water (and get unhappy thoughts) at first, but they'll get over that.
Bringing leather lets you make bags and leather armor, which are both fast, immediately useful and a nice savings. Bringing thread would let you weave cloth to make sacks and ropes - but that might be less critical and more involved than some other DIY items. Just because you "can" doesn't mean you should, or really will have time to - but it's always an option!
Bottom line, any finished good can be made if you've brought the raw materials, but most of them are not essential like the above, and thus you can generally wait until you find suitable resources on site or buy them from caravans.
 Biome considerations: Dude, where's my wood?
Some environments have a shortage of trees. While you can direct production of a lot of item types to other materials, beds and some items still require wood. In addition, it is difficult to make bins out of non-wood materials early in the game, especially without ready magma (since otherwise you'll probably need to burn wood to make metal bins). If you have an aquifer it can be even worse - stone may be difficult or impossible to access easily. While you can ultimately ask for wood from your liaison and buy whatever the humans and elves happen to bring, and eventually you can create a tree farm underground, tight wood will limit storage and sleeping arrangements for at least the first year if not longer. You may wish to plan accordingly if embarking in a site with sparse or no trees.
 Items for moods
When a dwarf is taken by a strange mood, he often needs obscure material or he will go insane and die, possibly with severe consequences to an entire fortress. Bringing along some of the harder-to-find ores (cassiterite, sphalerite, bismuthinite, garnierite), and putting those aside, forbidding their use "just in case", is spending a few points on an insurance policy. Many players also choose to bring a few items like pig tail cloth and cave spider silk just in case.
Alternately, if you're otherwise being minimalistic on gear you're bringing you can choose to bring a few valuable components to try to maximize the value of mood items. That artifact animal trap will be worth a lot more if your woodcarver grabs a blue diamond instead of moss agate.
 Free Equipment
The only free equipment you get in Dwarf Fortress mode is the wagon (which can be broken down for some logs), its draft animals, and your dwarves' clothes. Unfortunately, your dwarves don't get any free equipment based on their skills. This is in contrast to Adventure mode, in which the only equipment available on starting is free equipment.
That said, it is possible to procure additional containers by choosing to embark with a diverse range of items, ensuring that each is given their own bag, barrel or bin for less of the cost of the item by itself. See Container Mechanics and Free Items.
 Site considerations
Each fortress location offers particular challenges and opportunities, and can make different demands on your starting build. Your starting build may need to be adjusted depending on the region your fort occupies, the specific vision you have of your fortress, and what it will take to stay alive where you're going!
 General Surroundings
Simply put, if your surroundings are evil or savage, your dwarves have a higher risk of suddenly facing personal combat before they are safely behind their defenses. Consider bringing extra weaponry, in the form of axes, picks or crossbows. Hand in hand with those, consider skill mixes that include axedwarf, mining (the skill used to wield a pick), and marksdwarf (for crossbows).
The same is true if you are embarking near an exposed magma vent or an open chasm - these features can be seen on the embark map, but it's impossible to tell if they are "open" to the surface or not, until you are there in person.
Be sure to include some source of water on the map, preferably running water. Water is (almost) essential for any fortress. In cold and freezing climates, streams and lakes will often be frozen year-round and your dwarves may quickly die of exposure. In hot climates, murky pools will dry up, and in dry ones, rain will only rarely refill them, if ever. Choose temperate or tropical zones for an easier game.
If an aquifer is present in the first soil or stone layers (visible on the pre-embark menu), it may bar all access to stone and ore until you find a way through the water barrier. Consider bringing some stone for building, and ore for your first basic needs. This may be critical to getting your fortress running smoothly.
Mountains often have abundant ores, but at the loss of trees and plants. In previous versions lacking caverns, this was a serious drawback. Brave pioneers can dig down into the caverns to find essentials like water, mud, and plants. However, players should be aware that above-ground crops will not grow in mountain biomes, no matter how muddy you may make the surface.
Flatlands with at least some trees and gatherable plants can also make for highly successful fortresses. Advantages over mountain zones include abundant trees and plants and (unless frozen) more abundant water. There are even (rare) magma vents. More water also means a high likelihood of an aquifer being present. Make sure to check on embark.
The greatest disadvantage is the potential lack of exposed stone to mine. The first level(s) below the surface is often soil of some type, which offers no building materials. However, soil is mined much more quickly than stone (x3-x4 faster), and expansive accommodations (rooms) can be achieved quickly even by untrained miners. You will find stone, you just have to go down a bit for it - but that's what dwarves do, isn't it?
Training a Miner from No Skill to Proficient takes less than a month (~20 days with hauling disabled) in soil, and to Legendary in just under a season after. From embark, this means you should have legendary miners in early summer if you dig only in soil. Times increase slightly for each additional miner used.
With many features in common with some of the above locations, beaches are often a mix of ease intermingled with bouts of extreme difficulty. Minerals and trees are often abundant, as well as farmland and sand, but there is often no drinking water unless the biome has flowing water of some sort.
By definition, the settlement will fall between (at least) two biomes (one land, one water), potentially hazardous if the player expects a peaceful oceanside meadow, without realizing the terrifying ocean is full of amphibious zombie whales.
Hunters should be replaced with fisherdwarves and a fish cleaner (although the latter can be easily trained). Depending how much water vs. land, more starting wood and ores might be helpful. Swimming is rarely useful in Fortress mode, even at the beach, and can be trained.
 Desert, Glaciers, and Barren
Treeless (or near-treeless) biomes are challenging sites for a fortress: you get most of the disadvantages of a flatland site without having access to nearly as many trees and plants. However, near-lifeless zones such as glaciers are wonderful for players with slower computers, as there's little to burden the CPU but your dwarves and livestock. Deserts and barren areas often have sand; with a sufficient source of energy (preferably magma or coke, keeping in mind the scarcity of trees in these biomes), you can build almost anything out of unlimited glass.
 Technical tricks/modifications
Starting builds are located in data/init/embark_profiles.txt. They are editable as well as (usually) transferable.