Checking Index Accuracy: Searching for the Perfect Index: Part 1

We all would like to turn in 100% accurate, perfect indexes each time. How reasonable is this desire? Even if indexing were 100% science instead of the mix of art and science that it is, as Joe E. Bishop said at the end of Some Like It Hot, “Well, nobody’s perfect.”

What then is a reasonable amount of checking—proofing—to do after finishing input and editing of index copy? There has to be a tradeoff between one extreme of going through every single text page checking each entry and not checking anything at all.

Early in one’s career, the approach probably tended towards the high end of checking every entry. It is simple to sort an index in page number order (at least it is in Cindex and I would assume the same for Sky and Macrex). Then it’s just a matter of spending the time to verify each entry. This still will not guarantee a perfect index, if such a thing really exists.

What is needed is some way to calculate the minimum amount of checking to do to maximize the likelihood of catching errors. To rule out simple human errors, this needs to go beyond the suggestions for copy-editing that the Chicago Manual of Style (18.136) lists. The check has to be more involved that simple spell-check software. There are other tests suggested for editors, such as checking the first chapter, or the first x number of pages in a book against the index or checking every nth entry in the index. None of these methods can give any real guarantee of finding errors.

Statistical methods may offer some guidelines here. What might work is a simple calculation that will tell an indexer how many entries need checking to suggest with a pre-selected level of certainty the unlikelihood of egregious errors in the index such as wrong page numbers.

Only four things are needed for this calculation. First is an understanding of the term “random” in statistics. Second, the indexer will need a way to generate random numbers. A spreadsheet such as Excel will do this. However, within the next week I will post a link to a random number generator program written especially for indexers. The final two requirements have to do with the ideas of Confidence Level and Confidence Interval.

The next posting in this new series will cover the above items.

Good Reminder about Backups

The article linked below from the Small Business Computing Newsletter may be preaching to the choir, but it does have a good explanation of the different types of backups. There are also some suggested programs for backup.

A Beginner's Guide To Data Backup
By Sally Curran, November 28, 2007

Still Thinking about Moving to Vista?

Here's an interesting article comparing XP to Vista computer operation speeds:

Windows XP outshines Vista in benchmarking test

"New tests have revealed that Windows XP with the beta Service Pack 3 has twice the performance of Vista, even with its long-awaited Service Pack 1."

On the other hand, I do like the Sidebars in Vista, particularly some gadgets I've written for special purposes.

The Economics of Indexing: Part 7

This last essay will try to wrap up the previous six, hopefully with some reasonable conclusions. After that, please feel free to chime in with comments. Above all, I want to caution every reader that every number used in these articles except for those drawn from cited sources is interpolated using mathematical relationships, statistical functions, and some assumptions. I use specific numbers to make the cases more meaningful—“104” is easier to visualize than “a small number.” By working with statistical operations, I am trying to get at some underlying principles of the indexing business. The estimated numbers make it easier for me at least to visualize these principles. Some of the conclusions may well seem obvious, but at least after going through this exercise, I feel that there are some valid underpinnings to “common knowledge.”

Part Six worked with a hypothetical distribution of 236 publishers issuing an average of 68 non-fiction titles a year which needed an index. A larger number of smaller publishers (904) theoretically put out an average of four titles a year with an index.

From earlier chapters in this series, I have assumed there were about 1,300 working professional indexers in the U.S. Of these, I think it is probable that only a small percentage, say 10-15%, are employed, most likely in the journal and database publishing industries. This removes them from the pool of indexers available to take on individual books.

Based on other reasonable assumptions such as the 80:20 (Pareto) rule, supported by figures from the 2004 ASI Salary Survey Executive Summary, which notes that 8% of indexers earn over $80,000 a year (including employed indexers), and statistics from book publishing reports, I come to the following conclusions. Note: Journal indexing as well as alternative sources of indexing work and the timing effects of publishing, i.e., Spring and Fall lists, are excluded from these calculations.

1) Indexer income and numbers of titles indexed each year is represented by a Normal curve (so-called Bell curve; see link), with a very small number of people (104) indexing an average of 68 titles a year, a bit larger group (325) indexing nine titles a year, and a large number (871) of people getting the rest of the jobs.

2) Besides the numbers of titles indexed each year, to account for income disparity at the different levels, it is probable the higher earning indexers charge approximately 10% more per page than the lowest earning indexers. Intuitively this seems likely since the higher earning people probably have more experience.

3) The approximately 1,100 indexers who are not in the high earning group have about half the chance of high-end earners of getting a particular book to do in a given month. The part-time indexers odds are less than 25%.

4) Indexing is not a trivial business, since in the aggregate, the amounts paid (or in time spent by authors indexing their own work) probably are more than $20 million/year.

Does Ranganathan Offer Anything for Indexers?

Ranganathan who? I hear the non-librarian indexers cry!

Shihali Ramamrita Ranganathan (1892-1972) was a librarian and mathematician. I believe he indeed offers several significant thoughts for indexers. The first and perhaps most important for every indexer to remember is his Fourth Law of Library Science—Save the Time of the Reader.

This applies in so many areas in writing a good index. I’m sure readers can think of numerous ways this can be done—for example, double-posting references instead of using See references when there are only a few page locators. Nancy Mulvany in Indexing Books (p. 105) gives this example to replace a cross reference “cars. See autos.”

autos, 54, 89
cars, 54, 89

For indexers, going beyond the simplistic but so important Fourth Law, it could be helpful to spend a bit of time with Ranganathan’s major contribution to library science, his “faceted classification.” This is an alternative materials classification system that can be applied to information.

Faceted classification, also called colon classification, uses a combination of letters and numbers to describe a document's contents. With the additional of some philosophical terminology with links to Hindu thought, it is a five-dimensional approach that looks identical to advice given newspaper reporters in writing good articles:

1. Who: Personality (what is the main topic?)
2. What: Matter (what is it made of?
3. How: Energy (what happens with the object?)
4. Where: Space (where is it?)
5. When: Time (when is it?)

In Ranganathan’s own words, the approach to classifying any material is:

1. Study the context, the content, and the users
2. Identify the major types of concepts (facets) covered by the above
3. Use a taxonomic structure to fill each facet with concepts (isolates)
4. Categorize each document according to each facet

[Source: www.slais.ubc.ca/courses/libr517/03-04-wt2/projects/ranganathan/class.htm]

Supposedly, Ranganathan’s faceted classification led to Yahoo and other Web systems.

Delving too deeply into the actual specifics of the Colon Classification system is likely to produce a headache, but to give a brief idea, here is probably the most frequently cited example, which appears in too many different Web pages to cite: “Topic: research in the cure of tuberculosis of lungs by x-ray conducted in India in 1950.” Call number: L,45;421:6;253:f.44'N5. Components: Medicine, Lungs; Tuberculosis: Treatment; x-ray: Research. India'1950. To pursue this whole concept, which has implications for Web design and information architecture, search “colon classification” or Ranganathan in a search engine?

However, where this post started was with insights for working indexers from S.R. Ranganathan. So here is the bottom line: If struggling with a difficult passage in a book and pondering how to create index terms, would it help to analyze the material in terms of the five elements of faceted classification—Who, What, How, Where, and When?

Comments are welcome!

The Economics of Indexing: Part 6

Please do not decide publishing is an extremely lucrative field after the example towards the end of Part 5. That was a theoretical case of one very successful book. According to industry figures, 70% of books lose money. In actuality, the book publishing margin is less than 20% and often under 10%. Publishers Weekly 8/30/2004 includes a table of 2003 margins. For example, Houghton Mifflin's was 3.9% and John Wiley's 14%.

Selling out a full print run like the example has to balance seven or more books that barely break even or lose money. The significant fact to take away at this point is the publishers, perhaps excepting huge conglomerates, do not have a lot of money to play around with, so this impacts all decision areas at the corporate end, including how much is paid to an indexer (if the press is picking up the cost), or, how much the editor can get the author to pay for an index. The index price therefore is much more controlled than a freelance indexer, sitting at home and feeling that a $4.25/page rate is justified, may imagine.

This brings us to price elasticity of demand. While classical economists probably would feel I am using such concepts a bit loosely, I do think at least the broad ideas can shed some light on the indexing business. I believe thirty or more years ago the price for indexing was highly inelastic for two reasons. First, the number of indexers was considerably smaller. Second, before the Internet and Web, market knowledge of editors and authors was limited. It was not easy to find a new indexer. Clients either had a short list on a Rolodex or got names via word-of-mouth. I know my business built from 1972 for many years primarily through editors asking other editors, even at other presses, for names.

In Chicago in the Seventies early attempts to form a group at a dinner typically brought together less than ten of whom at least half were wanna-be’s. Even in the early Eighties, there were only two of us in that area using computers to index. Probably most indexers were concentrated in New York. Therefore, competition played a very small part in getting indexing work. Prices for indexing were generally set by the publisher—from one I was paid $4.00/hour and thought that was pretty good. There was some variability working directly with an academic author but still the price needed to be pretty low or the author would simply turn to a graduate student.

Today, the presence of the Web has increased knowledge of the indexing market by both editors and authors. The large increase in numbers of indexers over the past ten years enlarges competition possibilities. The nearly daily or at least weekly introductions on the email lists from another new indexing student raise some interesting questions about the future as well. Survey results for an article I wrote on Suite101.com, "Indexing as an Ideal Home Business” suggested between 82% and 94% of working indexers rely on email and/or the Web for marketing. This is good for indexers, but it also means editors and authors easily can query a number of indexers for price quotes.

What effect will these changes have on price elasticity of demand and supply—now and in the future? Others have commented on the low importance of an index from a corporate bottom-line sense. Yes, indexes mean more sales to libraries, but sales to libraries are in some ways counter-profitable for publishers. It’s a case of one sale/many readers (libraries) or many sales/many readers (bookstores). Obviously the latter increases sales revenue. Publishing developments such as the increased use of packagers adds more hands dipping into the pie and further complicates the economic equations.

I have a rather pessimistic feeling that indexing prices are becoming more elastic, that is, as higher prices are charged, several things may change. In one scenario, demand may drop, in the sense of either publishers focusing on profit margin and cutting out some indexes or authors deciding they can live without an index. In another possibility, increasing the supply of indexers increases competition and market knowledge of buyers through the Web may impact index prices negatively.

Next (and last in this series), I want to look at how my hypothetical 236 publishers with an average of 68 non-fiction titles a year and 944 publishers averaging four titles a year works out in terms of predicting how many of the available number of indexers will get assignments.

Craigslist for Indexing Work?

I would never have thought of looking on http://www.craigslist.com/ for indexers, but the pages turned up on a "Monster" list of freelancing work from the FreelanceSwitch.com site
Here's an example, although it's for a full-time position with a legal publisher in Oakland, CA.
http://sfbay.craigslist.org/eby/wri/459416253.html

===========
This FreelanceSwitch site has a lot of interesting stuff, so I'll be skimming it regularly to keep you updated. For example, "13 Breeds of Freelancer and How to Up Your Game," is presumably tongue-in-cheek, but there are bits and pieces that do resonate.



http://freelanceswitch.com/clients/the-13-breeds-of-freelancer-and-how-to-up-your-game/

The Economics of Indexing Part 5

Regarding my “house of cards” comment in Part 4, please keep in mind what I’m trying to build here is a thought experiment, not an exact scientific statement. The point of these essays is to try to discover some of the underlying principles of the indexing business. It is not, for example, to determine with statistical exactitude how many publishers are hiring indexers. So when at the end of Part 4 I refer to 236 publishers with an average of 68 non-fiction titles and 944 publishers averaging four titles a year, of course I am not maintaining those are exact numbers. They are used for a working hypothesis, which is that a small number of publishers put out a lot of non-fiction titles each year, and a larger number put out a few—which ought to be intuitively obvious. Other statistical services seem to provide similar numbers. See for example: Publishers with 10 active ISBN identifiers: 73,000; Publishers with 11-199 active ISBN identifiers: 11,837; Publishers with 200 or more active ISBN identifiers: 1,804. --PMA Newsletter, September 2003 http://www.PMA-online.org and http://tinyurl.com/8qlcs (www.parapublishing.com) with additional data.

With that proviso, picking up with the last paragraph of Part 4, what can we discover about how the concentration of publishing might affect how indexers are chosen (keeping in mind at this point, I’m only focusing on publishing houses—authors hiring indexers for self-publication will have to come later.

To analyze the question, I’m going to make a huge assumption—that the market for indexers is rational and obeys at least some laws of the marketplace. Some concepts that will be used are “Price Elasticity of Demand” (PED), “Fair Value,” “Market Price,” and “Competition.” Readers not familiar with the meanings of these terms in an economic sense can find reasonably accurate explanations of all through Wikipedia (http://en.wikipedia.org) .

Demand can be either price elastic or price inelastic. In general terms, if elastic, when the price increases, the demand will drop and vice versa. If inelastic, the demand doesn’t change very much no matter what the price changes. Compare the demand for chocolate, which is elastic, versus the demand for water, where the demand is inelastic.

Generally, PED deals with goods. However, services also can fall under this concept. On a simplistic level, goods and services seem two different things, but in reality the two merge together. Indexing is both a good and a service. Indexers supply a service (indexing the book), but the product delivered to the ultimate consumer is both a service (finding material in a book or journal), and a good (discrete value added to the printed material). How does this idea of price elasticity effect things like fair value, market price, and competition?

Let’s take those hypothetical 944 publishers with an average four non-fiction titles a year. For discussion, let’s create one example using more statistics from the parapublishing.com Tiny URL several paragraphs above. A trade book published by a small press sells for $15.77 (average price in 2002). The print run is 1,500 copies. The book costs $10,388 to produce. It goes to bookstores for 45% of the sale price and sells all 1,500 copies (extremely unlikely). Net sales for the publisher are $13,000. The book costs $2,700 to produce (typical production costs from the same URL). The publisher pays 7.5% royalties and ends with a net profit of $9,413 while the author earns $975.

Refer back to this last example in the next post in this series when we get into demand elasticity, value, price, and competition. (I’m trying to keep these posts to a readable, i.e., around 600 words or less in size.)

Networking Suggestions

The linked article suggests using social networking and recommends some particular ones. I'm not endorsing these, but they might be worth checking out. If the lengthy link is broken on two lines, try the tiny url version.

http://getdirectclient.blogspot.com/2007/11/why-and-how-to-network.html

or

http://tinyurl.com/35ts9r

The Economics of Indexing: Part 4

Moving into some speculations mentioned at the end of Part 3, is there any way to get a handle on the maximum number of indexers publishers are likely to select? If so, what can be said about this number in comparison to our estimated pool of indexers? The logical next question after these would be how many more indexers (if any) can the field absorb and still guarantee sufficient work for all? In response to a comment on Part 3, I’ll try to add periodical/journal/magazine indexing to the mix.

Perhaps the Pareto effect, also known as Pareto’s Law, can shed some light on the first question above. The Pareto effect postulates that a small proportion of causes produce a large proportion of effects. It is commonly known as the 80:20 rule. For example, in measuring library circulation (checkouts), it often can be demonstrated that 80% of the use of the book collection comes from 20% of the holdings.

Looking at this in indexing, an indexer may have a number of clients in a year, but does 80% of the work come from 20% of the clients? This is certainly a common principle in other businesses. More importantly, does it apply to indexers getting work? Do 20% of the working indexers get 80% of the jobs? If one aggregates data from the ASI 2004 Salary Survey Summary, including the 11% earning over $100,000 a year and the 25% earning less than $10,000 a year, it would seem an approximation of the Pareto effect is working here as well. Either that or the high earning indexers are charging far higher fees while the lower earning indexers are charging much less, which seems intuitively unlikely. If Pareto holds true for getting indexing jobs, how does this apply up the line, that is, in how publishers select indexers?

Going back to publishing firms, out of the 81,000 publishers in the U.S. the Pareto effect probably applies here as well. Twenty percent of the publishers may put out 80% of the titles. Actually, as the publishing industry has moved to a more monolithic structure and small presses have folded, the concentration among large publishers goes beyond the 80:20 rule. According to Publishers Weekly (4/25/2005), the top five publishers earned half of the sales in 2004. Although sales figures and number of titles published is not the same thing, it seems safe to conclude at least 80% of non-fiction titles are published by 20% (16,200) of the publishers.

In reality, the actual number of publishers we’re considering is much less than this, probably on the order of one-tenth. For example, limiting a search on www.writersmarket.com to non-fiction book publishers results in fifty-nine pages of names with twenty per page, or 1,180 publishers. Based on my experience, this is much more likely to be the order of publishers hiring indexers on some sort of regular basis. And probably only 20% of these publish more than a few books a year. Skipping some steps here for brevity (if anyone is interested I could share some of the spreadsheet calculations I’m using here), and working with the Writers Market figures, I come up with around 236 publishers doing an average of 68 non-fiction titles/year and 944 putting out an average of four titles. In looking cursorily at individual publisher output on Writers Market.com this isn’t that far off if median title output is considered.

What would this mean for indexing opportunities? That’s a subject for the next installment. The further I go into this, the more I fear I am building a house of cards. However, I find the speculation interesting and hope some readers will offer enlightening comments.

New Web Book Indexing Tool (not exactly)

Interesting idea at http://www.bookbump.com

Donation Opportunity

"There's no such thing as a free lunch."
The idea of a "free lunch" dates from the 1840s and moved from the West to the East, getting fancier as it approached Eastern bars and hotels. At some point, someone probably said something like, "Pay for that drink first before you get the free lunch."
More a propos to the current Indexing Students blog, the famous economist Vilfredo Pareto is supposed to have asked, "Can you tell me where I can get a good lunch, free?" The reply was, "There are no such restaurants." Pareto replied, "There is a general rule of economics for you!"
It is difficult to offer a weblog (blog) totally free as well--at least and have it amount to something.
Please consider making a donation to help support and keep the blog running.

Subject and Purpose of this Blog

This blog is not intended in any way to supplant the existing email lists or professional societies. It simply offers another way to gain and share information about the field of indexing.

All comments will be moderated before posting but only for relevance and acceptable language (no profanity, obscenity, potentially libelous statements, etc.) The purpose of this is to ensure that all comments remain on target and do not degenerate into time-wasting chatter as well as avoid spam comments.

Appropriate Subjects for this blog.
1. Comments on the freelancing experience, particularly as it applies to indexing, success stories, ups and downs of working at home.
2. Indexing Meeting Links: Alerts to upcoming regional and national indexing meetings; course offerings, podcasts, etc.
3. Marketing Tips: Marketing indexing services, leads, alternative sources for indexing jobs .
4. Newsfeeds: Publishing and technology related RSS feeds .
5. The Business of Indexing : Advice on calculating fees, dealing with difficult clients, practical questions and advice.
6. Using Software: Tips and tricks for indexing software, useful utilities and other programs for indexers.
7. Web resources: Suggested links to useful Web sites for indexers.

Indexing over the Years

My first PC was an IBM 8088 I bought in 1982. The speed was 4.77Mhz.The machine had 32KB of RAM (that is Kilobytes–not Megabytes). There was a jack to attach a tape cassette for backup. There were two 5-1/4″ floppy drives, and I had to buy an external hard drive with the magnificent capacity of five megabytes. Within two weeks, I traded that drive in for a ten megabyte drive. All told, including monitor and printer, the cost was nearly $8,000.

The indexing software I used was a database program called <> written in the Forth programming language. However, after ten years of indexing on 3X5 cards and typing out the index on an IBM Selectric typewriter, I thought I was on Cloud Nine. One major contract that I had for many years involved an annual index to the proceedings of a government agency which ran to approximately 7,000 entries. Visualize that being done on 3X5 cards! It filled seven card file boxes (like shoe boxes but sturdier), and as I finished each annual index, I got more and more worried about possible loss. How do you backup 3×5 index cards?

With the first computer, I cut the work time for that big project from 300 hours in half. In later years, using new indexing software, keyboard macros, and voice recognition, by the time I passed the project on to a mentoree, the original 300 hours work had been reduced to about sixty hours.

Indexes in those early years were delivered by mail or messenger service. The next big development was being able to transmit an index file electronically to the printer because the publisher did not have the capability to receive it via FTP (file transfer protocol), which along with Telnet, preceded the Web.
Those entering the indexing field now should be extremely grateful about the developments over the past thirty some odd years. Productivity has increased enormously as turnaround times have diminished. Indexers often were allowed a month or more to complete an index using cards and a typewriter whereas now a week or less is typical. The net effect of course, has been to increase the potential for income. However, given rises in the Consumer Price Index, it might be interesting to compare actual spending power between those early days and today.

Indexing Meeting Notices

Gale Rhoades Says: Here are the known dates for 2008 Macrex (North America) workshops: March 13th (Thursday) meet for dinner 14th (Friday) Play Day 15th & 16th (Saturday & Sunday days) workshop 15th (Saturday 5:30–­7:30p.m.) The Gathering (open & free to all; please RSVP) 17th (Monday) Voice Recogntion (Vista and Dragon NaturallySpeaking) October 16th (Thursday) meet for dinner 17th (Friday) Play Day 18th & 19th (Saturday & Sunday days) workshop 18th (Saturday 5:30–7:30p.m.) The Gathering (open & free to all; please RSVP) These events will all be held at the Best Western Grosvenor just north of the San Francisco International Airport. If you are interested, please request a brochure when it is available — just email macrex@wisebytes.biz

Alternative Sources for Indexing Jobs

In case you missed this article on suite101.com:

“Marketing Your Indexing Services: Finding Potential Alternative Clients for Indexing Work,”

Hourly Rate Calculator

This calculator seems to take into account everything but the kitchen table. However, for an indexing student considering the business, it may provide some helpful information.
http://freelanceswitch.com/rates/

The Economics of Indexing: Part 1

In the enthusiasm of studying for a new career such as indexing, it’s easy to get carried away with income potentials. The reality is that unless you are very lucky right from the start, it may be years, if ever, before you can support yourself by indexing. On the other hand, if your need from this new business is just some supplementary cash, then success in the field is easier to come by. What follows is not meant to discourage any new indexing student. It simply reflects some realities I perceive after 35+ years in the field.

Part of the income picture is how long you work on indexing. Most experienced indexers I know feel it is hard to turn out quality indexes working more than 4-5 hours a day on the index itself—excluding all the other parts of running a business. Indexing is not only mentally exhausting; it can be physically tiring as well.

How much you get done in those indexing hours depends on your speed. In the early days, when still unsure about your skills, you may not be able to index more than 5-6 pages per hour. Later, speed should increase, unless the material is very abstruse, to at least ten pages per hour. For extremely simple material and with experience, you may index twenty or more pages per hour.

Starting out, if you index six pages per hour, and charge $3.50 per page, that equals a gross income of $21.00/hour. In a five hour working day that translates to $105/day; $525/week, and, assuming you could get enough work to index fifty weeks a year, a gross of $26,250 annually. With lots of experience and clients, working indexers now probably gross around $40-$60/hour.

A figure generally accepted in freelance indexing is that net income is about 50% of gross, after deducting various taxes, operating expenses and time spent on marketing, etc. So the gross above may be not much more than $13,125. The reality is that in most cases, it may take months to get the first job, and in the first year you may only get a few jobs. The first year net income then, even with increased speed, probably is going to be considerably less.

It follows, then, that to maximize income, one important factor is how to price your services, specifically, how to estimate how much to charge for a specific index. I’ll talk about this in another blog posting.

The Economics of Indexing: Part 2

This post and several that will follow are ones on which I really would like to see comments.

I was ruminating this morning on a question that indexing students probably ask themselves when they consider indexing as a profession: “How much room is there in the field for me?” Since there are only a few factors that influence this question, it seems an answer should not be that hard to find. Unfortunately, it is darn near impossible.

Certain figures are easy to come by. Yearly publishing output is easy, although always several years behind. Bowker, Inc., publisher of Books in Print, collects data from 81,000 U.S. publishers. Bowker compiles annual data on publishing. You can see a breakdown by subject areas for trade books and university press titles at http://www.bookwire.com/bookwire/trade.html and http://www.bookwire.com/bookwire/university.html
Besides totals, the individual subject area data might be interesting to students who talk about wanting to specialize in certain areas.

Working with 2004 data because 2005 trade publishing is not final, we find there were a total of 33,518 non-fiction titles put out by trade and university presses in the U.S. If we knew the number the number of working indexers and of these titles with indexes simple math should give a clue as to whether the field was getting saturated.

However, those latter two numbers are very shaky. ASI’s 2004 summary of its salary survey notes that approximately one-third of members responded to the questionnaire. Since there were 227 respondents, this implies a membership of about 688. This is down from around 1,000 when I was on the Board in 1999, which could be read in several different ways. For instance, it might support some calculations that follow regarding total number of working indexers, if people just left ASI but stayed in the field. Calculating the number of non-affiliated indexers moves us into the realm of crystal ball gazing.

It is possible, though to set up some thought experiments to run through various scenarios. For one hypothetical look, suppose there were an equal number of non-ASI indexers, with an opposite breakdown by full and part-timers. There’s some logic to this assumption. Part-time indexers, particularly when 25% of ASI members earn less than $10,000/year, might be more likely not to pay ASI membership dues. In this scenario, there might be around 1,300 indexers working in the U.S.—a figure that is close to one occasionally mentioned on the email lists of 1,200.

What percent of non-fiction books are indexed—and of these what percent are indexed by authors? In the first case, it seems reasonable to assume that a high percentage, perhaps as much as 90%, of university press books are indexed. In the trade area, let’s say only 50% are indexed. Massaging some numbers from the above sources suggests that perhaps 20,000 titles (2004) were indexed. Feeding in other percentages and mixes can lead to a low in the area of 8,000 to a high of 25,000 titles.

Let’s split the difference and work with 16,500 titles indexed. This number divided by 1,300 indexers would equal only 13 books/year per indexer. Since ASI’s survey claims 11% of full-time indexers earn over $100,000 (a figure I have a great deal of trouble believing), the median income of full-time indexers is $45,000-$49,999, and their average per page charge is $3.26-$3.50 per page, it’s obvious a small number of indexers are indexing a lot more than 13 books (or charging a lot more than $3.50/page) a year to earn the income reported. I suspect you would have to index at least 45 books a year to match that median income.

[More to follow]

The Economics of Indexing: Part 3

Carrying on with thoughts about the viability of a career in indexing, I want to look at ways to estimate how many indexers are required to index the annual output of non-fiction titles. Yesterday, I developed a scenario that hypothesized around 1,300 working indexers in the U.S. Broken down into full and part-timers, this worked out to 650 in each group. (Remember, these figures are very rough and based on some assumptions. I suspect the estimates are on the high side, but they are the best I have at this point.) It’s worth noting the full-time indexer figure does include employed as well as freelancer workers.

What is the maximum amount of work each of these indexers can do in a year? Remember, four to five hours a day inputting is probably a maximum for most jobs for freelancers. As full time employees, I suspect while they may be working a thirty-five or forty hour week, again, five hours is probably tops again because there are still other tasks and time-outs associated with the work, including editing, breaks, and lunch hours. So, let’s take five hours input as a typical day for both free and part-time workers. This means maximum possible indexing input time, while the real time will be much less because most freelancers are not lucky enough to have their work calendars completely filled fifty weeks a year.

To work on a maximum, let’s assume twelve pages an hour instead of ten. At five hours a day, three hundred pages a week, 15,000 pages a year (fifty weeks). I’ll use this figure for both full and part-timers for simplicity. So, theoretically, 1,300 indexers could turn out 19,500,000 pages a year. Working from the minimum and maximum number of non-fiction titles suggested in yesterday’s blog (8,000 to 25,000), and a hypothetical average of 325 pages per book, we have at the low end 3,480,000 pages and 8,125,000 at the high end. Under this theory, we don’t need 1,300 indexers—in a perfect world, between 232 and 542 indexers could handle the total printing output of non-fiction in the U.S. in a year, assuming scheduling of proofs were perfect as well.

Since the world is not perfect, page proof production schedules frequently slip, and we know that all indexers aren’t working all the time, the reality is that those three to eight million pages are being divided among far more than 542 indexers. The questions then become: 1) what is the maximum number of indexers that publishers are likely to select to do their indexing; 2) how does this compare to the number of indexers in the field; and 3) how many more indexers (students) can the field absorb and still guarantee sufficient work for all?

I’ll provide some thoughts on these questions in succeeding posts.