Arisia has a corporate (hereafter "continuing", for improved generality) and a convention treasurer. The continuing treasurer establishes accounts for the organization and with vendors such as USPS, Fedex, and Kinko's. The continuing treasurer handles acquisistion of durables like cash registers that will be needed in treasury. If we depreciated our other durables, the continuing treasurer would handle that. To the extent that the convention uses continuing accounts -- in our case, for merchant transactions and for artist escrow (see below) -- the continuing treasurer is responsible for the accounting, though it can be delegated. And the continuing treasurer established the two accounts that Arisia uses respectively for even- and odd-numbered conventions. Finally, the continuing treasurer provides oversight for the convention treasury.
The convention treasurer is responsible for everything else.
Convention treasury, to my mind, can be broken into four parts: a Treasurer who maintains the accounts; a Comptroller who maintains the budget; a Cashier who handles cash, credit cards, and checks taken in at-con; and an Auditor who looks for mistakes. Certain treasury functions are often nominally handled by departments -- constructing a reckoning for the cash-boxes, for instance -- but sometimes there is no reckoning except that made up by the Auditor. Certain other functions are nominally handled by treasury -- deposits from registration, for instance -- but can be farmed off to the departments. Finally, sometimes departments -- Sales and Insurance are common examples -- report to (and must be supervised and trained by) Treasury, but need not be *done* by Treasury if someone can be found to handle them.
If there is no continuing treasury, then the functions I ascribe to it above can be split between the four parts I describe. But the split will be unevenly to the "Treasurer" function. I don't really recommend trying to combine continuing and convention treasury if only because of the oversight issues it raises.
This breakup means that the Convention Treasury function can easily be split into two parts. My opinion is that a Treasurer/Cashier and a Comptroller/Auditor is the most sensible split, but you could split it some other way too. Four-way splits are also possible but seem to me to be a waste of resources for a typical mid-sized regional. If you split up treasury, be sure the chain of responsibility is left intact.
All told, I spent between 250 and 300 hours, and my staff spent a total of another 50 to 100 hours, each of the years I did treasury for Arisia. This is far, far less work than any other job of even remotely similar importance.
Treasury does not require a whole lot of space. For a mid-sized regional, a large table, a few cubic feet of storage space, and two or three chairs are the basic requirements. Treasury needs to continue operations later than most other functions at con, so putting it in Operations is problematic unless it clearly won't get in the way during load-out. A small hotel room, or an otherwise unused function space, are both acceptable. Proximity to functions is important, but so is proximity to the Staff Den. If you use a hotel room, don't forget to put out the Do Not Disturb sign! If you use function space, examine its security carefully. If the space isn't secure enough, you will have to keep the cash in the hotel safe whenever the space is not staffed, and this can be a pain.
I made up a schedule based on opening and closing times of various areas, pickup schedules to ensure that cashboxes wouldn't get too full (this depends on who's staffing the area but I aimed for $1000 pickups), vendor deliveries and pickups that might need checks, bank openings and closings, and other commitments on my time (Arisia had twice-daily meetings of what Minicon would call the "executive committee", and I had a panel commitment or two). I scheduled worksessions and bank drops into the holes in this schedule; this resulted in half a dozen worksessions, each of which lasted about an hour and involved two or three staff counting cash, preparing deposits, and doing preliminary audits of cashboxes. My schedule had me on site briefly on Thursday evening; back at the hotel midmorning Friday and busy until nearly midnight that night; on site Saturday morning until Saturday evening, and on site Sunday morning until Sunday evening. Monday after the convention was a holiday, so I didn't make the final bank drops until Tuesday; this only took a short time as the drops had been prepared and put in the hotel safe ahead of time. This schedule meant that I could commute to the convention without causing any problems, but I did have to be sure to keep to it, and I carried a pager just in case.
The first time I did cash-box handling, as treasury staff, we used a traditional central "bank" box, and made up deposits and withdrawals from that shortly before they were needed at the cashboxes. We had some standard forms for this that apparently the entire SF community uses for this purpose, though they had been modified somewhat and not for the better. This process was incredibly inefficient! Balancing the central cashbox was a pain, and when there were errors it had to be balanced by bill size. We had no way of planning how to stock the central box with small bills, or of determining whether we had enough small bills as the convention progressed. This meant we kept all the small bills in the box and only made deposit drops of the big bills; we wound up with way too many small bills at the end, which made balancing the box that much more of a pain. Counting out the cashbox setups, and finding more staff to verify the counts, also took a lot of time. I vowed never to do cash box handling this way ever again.
Since then, Arisia has had a policy of pre-building every cashbox setup. We made a little Excel spreadsheet of all the cashboxes, what bills and coins they needed, and how often they needed setups (we did every day rather than every cashier, since most areas only had one cashier anyway). We included a couple of spares in case we forgot something (though we also had cash out slips so we could build a new box out of a running box if we really needed to). Then, a few days before the con, I went to the bank with a check for $2000 made out to myself and a printout of the bill size summary and got to amuse the teller by asking her for a thousand dollars in $5s. The bank didn't have enough $2s so I got those amounts in $1s. I went home and counted out the setups, and had one of my staff confirm them, and sealed them in envelopes with the cashbox name and time on the outside. I brought these to the con and put them in treasury's cash-storage area. The one mistake I made was stocking cashboxes with full rolls of each coin denomination they would need. Half-rolls would have sufficed in most cases.
Since we had all the setups built, we never had to take out any cash from a pickup. Pickups were counted twice (who exactly did the counts depended on the department), merged together, and the merge counted only once. Then the merged pickups were sent to the bank and the teller provided the second count of the merge. To keep the audits simple we did not merge pickups from different days. We thus had a Friday afternoon drop, a Saturday morning drop that contained only cash from late Friday pickups, a Saturday afternoon drop, and two simultaneous drops on Tuesday morning for the Saturday night cash and the Sunday cash. Each of these drops was under $10,000, which was nice in terms of lowering our courier paranoia level.
Some banks will not accept more than a dollar or so in loose change. I solved this problem by personally purchasing the extra non-rolled coinage from the merge before depositing it. Having a small central bank with half a roll of each denomination of coin and maybe $10 in singles would have made this process easier, as I could have sold coin to the merge instead of buying it when the merge had more than half a roll. If treasury were in Ops, the Operations petty cash box could be used for this purpose.
Arisia has toyed with use of 50-cent and $1 pieces. They're fun, but people mistake them for quarters and for each other; this makes balancing the cashboxes harder. Maybe the Sacagawea coin will fix this. It's also hard to get wrappers for them, and it's hard to get the bank to take any quantity of them in a deposit, so I don't think they're good for more than occasional use. I see no argument against $2s, though I caution against promising them to people as the bank may not be able to deliver.
update 2/2003: Arisia now has an infinitude of 50-cent-piece and dollar-coin wrappers (and also currency straps). If you want some, just send email to treasurer@arisia.org.
Toward the end of the convention especially, as staff get tired, it becomes tempting to take shortcuts with cash-box procedure. Miscellaneous pickups and even cash pickups can be counted later; the bank is closed anyway so the time pressure is removed. The problem is that the time and sometimes even cash-box source information can be lost if a form is not at least partly filled out at pickup. I can think of two measures to reduce this temptation. First, a unified cash/misc. snap form will combat the urge to count only the cash, as there is a place for the misc. on the form. Second, stocking cash-boxes with snap forms -- perhaps even pre-filled out with the cash-box source identified -- rather than supplying them to couriers, will make it that much easier to fill out forms properly when performing a pickup.
About those forms: I recommend triplicate forms: for the department, the cashbox reckoning, and the bank deposit reckoning. I experimented with printing them on or attaching them to the outsides of envelopes, but I don't recommend this as it is a security risk. I strongly recommend 3up rather than 2up forms. My forms don't have explicit spaces for signing off on the count, which might be a bug in a larger operation but didn't cause problems for me.
An aside: Arisia is not unique in having the cash flow for prebuilt setups (see cash flow, below). We could pretty easily have done more frequent fresh setups or even one setup per pickup, especially since we could have used smaller setups in that case. I have a hard time believing the assertion that this method is impractical for a Worldcon. Setup-per-pickup would exacerbate the problem of figuring out how many setups you need but Worldcons operate when banks are open so you can make more without having to dip into your pickups.
Life is much easier if no coins are used at cash-boxes. If you need coins, keeping the coin denomination count down is a big win. Arisia is lucky in that the sales tax is 5% in Massachusetts. This means that whole-dollar sales come to even nickel amounts. Also, clothing is not taxed. By imputing the sales tax on the few non-clothes items at the T-shirt sales table, we can keep all sales to even dollar figures there. This is not practical at the art show, however, so the art show cash-boxes are stocked with nickels and quarters (no dimes, except for Sunday art show sales). Because the art show staff are not all from Massachusetts, a sales tax table and a calculator are both good ideas. The sales tax table only needs entries for whole-dollar amounts. Because of the nice round sales tax number, the one I made went to $100 with instructions to add $5 for each additional $100. This would have been less practical in a state with a non-integral percentage sales tax. I made my table include three columns: amount, tax, and total. This cut way down on (but didn't entirely eliminate) the number of adding errors.
I have only attempted to use cash registers at the T-shirt sales table. They made a big difference there, but only because I used a register which had item keys for every possible thing we were selling, clearly labeled. The advantages: vastly reduced error rate; summary printout at end of day. The disadvantages: programming the register is complicated and takes time, which means it should be done well ahead of the con, which means no changing your mind at the last minute what things are going to cost. Cash registers require power (but then so do credit card swipe-boxes). Cash registers break. If you have more than one area with cash registers, try to program them all so that they can be swapped around in this case.
I think that the print shop could use a cash register to simplify auditing at the end of the day. Here, however, the advantage is less. The problem is that prints and prices are not known far enough ahead of time to program them into the register. So while you can use PLU mode to require entry of the piece number (avoiding legibility problems on receipts) there is no mechanism for ensuring that the right price is paid.
I am reluctant to put cash registers anywhere that lines build up without having further experience with them in less-congested areas. This is not because cash registers make transactions take longer (quite the opposite), but because you now need another cash register to add another station, instead of merely needing another gopher.
Arisia has a credit card handling manual. This manual is FAR too complicated and graphically terrible and a better one should be made up. It would probably be best to have separate versions of this manual for the three typical credit card setups (kerchunker, swipebox, and printing swipebox) rather than having one manual that tries to cover it all.
Because of the many fallback possibilities for taking credit cards there is little reason to avoid their use. Swipe boxes run around $75 and printers around $40 on the used market. To my mind this means there is no point to not having as many printers as swipe boxes. They can also often be borrowed from other organizations, since reprogramming swipe boxes is simpler than reprogramming, say, cash registers. Be sure you have enough cabling, too! This is one of many potential pitfalls to last-minute equipment purchasing.
Credit card machines require phone lines, which can be very expensive in hotels. I would be very interested to hear the results of any attempts to use cellular phones instead. CC machines can share lines, but there is some debate as to the effect this has on their throughput. Finally, Internet-based credit card mechanisms show a lot of promise for at-con use, provided that the network infrastructure can be depended on.
Credit card machines do break and it is best to have someone responsible who is in charge of wrangling them.
The computer debate here is not as well settled as it is with registration. The clear advantage is that ambiguous or plain wrong forms can be identified in real time (for a line-of-fire system) or near-real-time (for a back-room system) and inquiries made to the art show staff. Though the improved responsiveness is an advantage, the at-con distraction to them, and the drain on at-con staff, is a clear disadvantage. If a back-room system is used, it may be better to do it in treasury's space rather than in the art show. If a front-line computer system is not used it is important that treasury get copies of the checkout sheets as well as the summary sheets and tally sheets. Copies of bid sheets would be nice too, but this is harder since they can't easily be done on NCR forms.
My experience with computers in art show specifically was that there were in fact many errors left over after the entry was done, and these had to be chased down in the usual fashion. It wound up taking about the same amount of time it would have otherwise. The reason I let it happen it that is I had someone who thought it would be fun to do, and who was willing to take the whole job if I let them do it. If the initiative to computerize had come from the art show, communication might have been better and this problem could have been ameliorated.
Regardless, the main treasury computer should not be used for data entry.
Financial software and departmentalized cash registers are the auditor's friend. Laptop computers are the comptroller's friend. Tape adding machines and photocopiers are the cashier's friend. Filing cabinets are the treasurer's friend. Do not get these confused.
Confused? I mean that the auditor doesn't need a laptop, and the comptroller doesn't need financial software (I used Excel for the comptroller function in '99). I mean that computers of any kind do not get you out of having to keep paper records and file them properly. Printed tapes of checks and charges are easier to refer to, and much easier to make, than entries in Quicken. And applying too much technology to a problem usually makes it harder, not easier.
For a long time the rule was that laptop computers dropped off the market after reaching the $1000 price level. This meant that, though used laptops capable of running anything Treasury might practically ever need could be had for a few hundred dollars, actually using them was a technical support nightmare. While this may change, so far the convenience of laptops vs. desktops has not been a compelling enough reason for a convention to buy one. Desktop computers may be worth buying if your comptroller or auditor doesn't have one, depending on many factors about your organization. Sure, desktops are a logistical problem. But the advantage to the con of treasury having any computers at-con is not worth money -- if you don't have relatively recent laptops you can borrow, leave all the computers at home.
What makes budgeting hard is the politics.
At first glance, budgeting is about figuring out what it will cost to do all the things the committee wants to do, and telling them whether they can afford it. The problem is that your job isn't over when you say "no" (or "yes, go think of more things to do"). Suddenly someone has to decide which things to cut, and which things to add, and where to be more extravagant, and which things to do more inexpensively. I'm in favor of emphasizing the latter option, but it often takes a lot of cajolery.
It's not even as simple as figuring out where the convention gets the most bang for its buck. To do things more cheaply, you must weigh the ability to find someone to do the work against the dollar cost, so you have to consider which areas have extra manpower. To cut things, you must figure out which areas' volunteers will leave if their event doesn't happen and which will work on something else, and consider this along with the relative importance of each event to the convention.
Any organization will have people who ask for way more money than they need and people who will ask for way less. It's useful to know which of these people are which. Although you should compensate for this in your budgeting, it's not necessary to do on paper where everyone can see it -- and in fact it can be counterproductive to do so. I have seen area heads who refuse to work unless I give them a $200 budget every year, which they never spend. Fine. Another area head always goes over whatever I give him by $500. These things balance out.
Most conventions don't know what all they're really going to do before they start spending money. So the budget has to evolve. There are lots of mechanisms for doing this. I like having a provisional budget, no more than 5% the size of the overall budget, covering activities that happen more than six months or so before the convention. This allows the con to spend money on flyers and parties and progress reports and so forth while figuring out what all is going to happen at the convention. Once things have gelled a little, this can be replaced by a real budget.
Some conventions have a discretionary, or "contingency", fund, perhaps 5% of the budget, which can be doled out by some responsible authority. I don't like this as I think it discourages fiscal responsibility. People are more likely to run over their budgets and seek forgiveness afterwards if they think there is spare money lying around. Some conventions have a 10% "margin of error" requirement between their income and their expenses. I found this to be useful in getting the committee used to the idea of fiscal responsibility. Now that they are, though, I don't think they need it any more.
Two tools I made heavy use of were unnamed line items and budget amendments. The differences between an unnamed line item and a contingency fund are first, that the item can't be added to existing line items, and second, that the item can be restricted to certain categories of unanticipated expense. The differences between a budget amendment and a contingency fund are first, that budget amendments require a more formal approval process, and second, that budget amendments can include subtractions as well as additions. The more formal approval process is good in that it keeps people from presuming that they can get the money, but bad in that you bring more business to the budget-approving body. One other nice thing about a formal budget amendment process is that it provides a nice artificial deadline (the meeting date of the approving body) for people to tell you what's going on in their areas.
I have occasionally used contingent budgets. That is, a formula for automatically adjusting certain budget numbers according to the pre-registration figures. I've also done this by hand, amending budget numbers when the estimate of convention attendance changed. The utility of this depends on how much uncertainty there is to your attendance.
One potential budgeting gotcha: many items' feasibility depends on the ability of the convention to borrow the equipment needed to do them. For some things we are willing to rent the equipment if necessary; for others we are not. This brings up two issues. First, how to budget. I prefer to budget as if we are able to borrow, and then amend if it turns out we are not. Doing it the other way, I think, leads to people thinking they are entitled to the unneeded rental money. Second, when it's worth continuing an event in the case of partial equipment availability. If the expenses slowly creep, it's important to have a threshold beyond which the event isn't worthwhile, and to pay close attention to see if it's going there. After all, if you're going to ditch something, it's a shame if you're already spent a bunch of money on it.
A law/chaos issue is the nomenclature of expense categories. Numbers for expenses are good because your paperwork is never ambiguous. This is especially good for things like use of a Fedex account, where the vendor can supply a short reference number but not a long description. The problem with numbers is that no one can remember them. I use hierarchical category names instead. For example: Con Chair/GoH Liaison/Travel. Quicken does not require that subcategory names be unique; there can be (and I had) multiple Postage categories. Keeping category names unique, or at least limiting their overlap, does clear up your paperwork, however.
Arisia, Inc. is doing pretty well financially and has the cash to be able to front money to the convention accounts. However, my experience has been that the convention does not actually need this money. We raise a few thousand dollars at the previous convention by selling memberships at a special year-ahead rate, and as long as that money winds up in the proper convention account we don't need any more front money. We do experience two dips, to about $4000 just before the pre-reg deadline and to about $2000 immediately before the con. This is with having to give the hotel a substantial deposit; in years that we haven't done deposits the dips have been much less deep. If your organization is in enough financial trouble that it uses the early membership money to pay for the previous con, then you might have to find some way to ease those cash crunches. If your organization doesn't have a lot of cash, you will need to figure out where money comes from if the convention runs at a loss (see "next year's membership money" above). But even if you run at a loss, cash flow will be good in the month after the con, and it will only be when the net-30 vendors and the artists have to be paid that you'll have to come up with the money.
Arisia gives out free memberships to the next years' con in exchange for working 12 hours on the current con. The current con is supposed to reimburse the next con for the incremental cost of a membership, which is fixed at $8 though I personally think the figure is between $5 and $6. But, since we get the year-ahead prereg money, this money is not required to make the cash flow work (and isn't included in the calculations I did above), and simply comping those memberships would be fine. If we didn't have this structure I suspect many of those people would buy memberships at the year-ahead rate, and our cash flow would be even better.
If you are starting a new convention (which means you don't get year-ahead sales), or your convention unexpectedly hemorrhages money, consider asking another convention-running organization for a loan or grant. Arisia has easily enough money in the bank to get a new convention, even a fairly large one, off the ground, and I hear Minicon is interested in funding new conventions as well.
The easiest way to save a ton of time is to delegate some amount of treasury work to the departments, particularly to preregistration. Give them a deposit stamp and some deposit slips and have them photocopy incoming checks and mail them in. Give them a credit card machine and have them process the incoming credit cards -- or put your credit card machine on the net and have them submit mailed-in credit card registrations there. The two keys here are communication and responsibility. You need to make sure you get copies of the checks. This is fine if registration already handles checks and passes them on to you. But if your con has someone else open mail and send checks directly to you without passing through registration, you may need to examine whether the reg people can be counted on to get you those photocopies promptly. Chances are you've really already done this in deciding not to have registration touch the checks in the first place. You also need to make sure checks aren't lost, cash doesn't get sent through the mail, deposit slips are filled out correctly, and deposits happen in a reasonable time-frame. Again, this is a trust issue with registration which is orthogonal to the question of whether to deputize the registration (or art show, or dealers room) staff.
I prefer Quicken 98 for Windows over earlier or later versions (later versions have less efficient screen real estate usage). Quicken and Microsoft Money (and all the other such things out there) are double-entry bookkeeping underneath. However, they generally cope poorly with book entries in more than two columns. And they make some assumptions about how you will use them which affect their suitability for running a convention.
The real reason Quicken is not marketed to small businesses is that it does not automatically handle payroll or complex depreciation. If you want to do these, you have to do them by hand. Most SF conventions don't bother with either of these (or even with simple depreciation, which Quicken can handle). Cons that do payroll (for baby-sitters, for instance) generally hire a service, so they don't have to remember to file W-2's nine months later when the convention's books are closed.
Quicken's attempt at multi-column entries ("splits") is sufficiently pathetic that I avoid it entirely. One problem comes when making detail reports; the register detail does not contain the splits information. Also, one entry is considered the "parent", limiting the potential complexity of a split. Finally, splits may not contain more than one entry in the same column (though this is a problem on paper too). This is particularly a problem for deposits of multiple checks from the same income category, for which it would be desirable to have a single entry in the checkbook register, for ease of balancing, but multiple entries, possibly under different names, in the income category detail. It's also a problem if I write a single check to a vendor for multiple items in the same category but would like to track the cost of each item separately in the detail. So, if I write a check or deposit to cover more than once category of income or expense, I make it to accounts receivable or accounts payable, and then enter the different lines there. A special "splits" account would work for this as well. I am told that "real" accounting software addresses these issues but my workaround works fine and their learning curves are steep.
Quicken has a whole bunch of register types (which it calls "account types"). The main differences between these are whether they are set up to handle assets or liabilities, and whether you can reconcile them. For some unknown reason only the "checking" (asset) and "credit card" (liability) types can be reconciled. So, you should use these account types for everything, from cash-boxes to accounts receivable.
Quicken's budgeting function is set up for continuing operations, not one-shot conventions. It defaults to amortizing your budget over the course of a year. This is a pain when you generate reports, because it will pro-rate all your line items by the duration of the report. You can make sure your reports go for a full year, but if your convention's books stay open longer than that this trick doesn't work. A minor hack is to pick a month and tell Quicken to put your entire budget in that month. Best is to use the month immediately prior to the first month in which your books have activity (though months shortly before that are OK too and there is some convenience to being able to make it January). Then you can make your reports go for 23 months before running into problems. But if your books stay open for 24 months you're in trouble, and now you have to go back and divide your budget in two and make sure all your reports go for two years (which isn't a standard option). And so on. This isn't a critical problem for Arisia, whose books open a year or less before the con and close within six months afterwards, but it'd be a big problem for a Worldcon.
Ideally a convention should have linked merchant (credit card), checking, and perhaps savings accounts. A convention credit card is also very handy. The checking account should be at a bank that has a branch, preferably a long-hours branch, near your convention site so you can make deposits. But there very well might not be any such banks that are willing to hold your merchant account. Oh, well -- a convenient branch is more important than a linked merchant account, especially since your merchant activity pattern will probably make it unnecessary to have more than one transfer per year from the merchant account to your checking account.
Do not get one of those enormous, full-page 3-up stub checkbooks. In fact, do not get stub checks at all. Duplicate checks are fine. Large-format is nice because it impresses people and allows things like multiple signature lines and expiration text, but whatever size you use, a 1-up, 25-checks-per-pack checkbook is much easier to deal with. You don't need the register-like qualities of the stub books if you're using Quicken anyway.
Arisia now has a separate account for writing checks to the artists. This isn't so much because we need one for escrow, as because we have a checking account with our inconveniently-located merchant bank. We transfer the merchant income (which exceeds the artists' sales) to that, pay the artists, and then transfer the rest to the main checking account (or directly to the corporation if we make a big enough profit). I wouldn't say that doing this is worth paying account maintenance fees, or even forgoing interest on money needed to keep a minimum balance, but the separate account is worth it to avoid wire charges, and having checks printed up that say "Artist Escrow Account" or something like that on them does keep the artists happy.
Whether to require multiple signatures is to my mind an open issue. Arisia requires it for amounts over $500, a fact which I use mainly as a stalling tactic if I want to investigate an expense before paying it. If you separate the treasury and comptroller functions it can be a way to indicate the comptroller's approval of an expense. You can also get checks that say "two signatures required" on them for accounts set up to really require only one. Whatever you do, note that banks cash checks with not enough signatures on them all the time, so don't imagine that it buys you any safety.
A trick: if a credit card is declined, try it again in a week or two rather than trying to get a new number from the card owner. This trick works with checks too, but check to see if your bank already does it for you automatically, and think a little about the bounce fees.