An Interview with John Morse, Part 3

Hosted by Emily Brewster and Peter Sokolowski.

Produced in collaboration with New England Public Media, with much gratitude to John Voci.

Download the episode here.

Transcript

John Morse: One of the great things, and I think this is sort of the next part of this story, is that by going online, we now know what are the words that are most frequently looked up in the dictionary.

Emily Brewster: Welcome to Word Matters. I'm Emily Brewster. Today, Peter Sokolowski and I continue our conversation with longtime Merriam-Webster publisher, John Morse. The process of converting an enormous dictionary that was set in lead type for printing on paper into an online dictionary was neither easy nor obvious. The line from a heavy doorstop of a book to a smartphone is not a straight one. In today's episode, John explains the process of moving from print to digital in an industry that had to change quickly from seeming very old-fashioned to being cutting edge in order to stay relevant and survive as a business. We think you'll enjoy this conversation as much as we did.

Peter Sokolowski: Now we're getting to the decision to go online, and the AOL was almost like a rehearsal.

John Morse: In retrospect, I think it served as that. I would say more than practice, I think it really was a goad because it really showed us that a dictionary website could be successful, and successful in a way that was meaningful to us, which is that it not only would provide a dictionary and get whatever traffic a dictionary got, but could be a brand building experience as well. We talked about this in the last session, but the reason why there is a Merriam-Webster today and the reason why we've had a dictionary publishing tradition that is probably longer, if we go back to 1843 or 1828, all the way up to current is probably one of the longest-standing dictionary publishing traditions you're going to be able to point to, it's because everyone involved in that has deeply understood the importance of branding, that branding is the thing you do that distinguishes closely similar products. And I think we understood that, and almost everything Merriam-Webster has done, whether it consciously or unconsciously, I think, has gone to build on branding. What the AOL experience let us see is that online publishing could be a brand building exercise as well as a purely monetary exercise. It would continue to reinforce the notion that Merriam-Webster is relevant. If you have a whole new generation of people growing up, whether or not the print dictionary is a perfectly adequate way of getting your information or not, still just feel like the right way to get that information is online. Whether it's a better experience or a more informative experience, it doesn't matter. That's just the way you get that kind of information. We really needed to make clear that, yes, that's fine, that's the way you like it, that's the way you get it, that we needed to retain that sense of our relevance, and I think that was as much the thinking as any, and I think the AOL helped us see that that was a possibility.

Emily Brewster: And the association with AOL was a significant one at the time. There were very few websites that people knew. AOL was one of the seminal websites. Everybody went to AOL.

John Morse: Yes, there were very few websites, there were still not that many online users. The interesting thing is in fact, AOL had some stiff competition. Prodigy and CompuServe were offering very similar offerings. And in fact, I think at the time when we were making our agreement with AOL, they may have had more subscribers, but clear that, the decision was a good one and because AOL survived the demise of CompuServe and Prodigy and was really one of our most valued licensing partners, second only to Franklin in terms of the importance to our business. Because of that, the signs were all around us that this is something we should be doing, and I think particularly the hypertext Webster and the number of people we knew were using unauthorized versions of the Seventh Collegiate was enough for us to understand that we really needed to get into this. Now, we sort of backed into it in an awkward, inelegant way. We started experimenting with, what would an online dictionary look like almost as soon as the Tenth was out. By '94 and '95, we were futzing around with how could you make an online dictionary. And references to it at the time say the public access dictionary. We kind of knew before we even had it launched that this would be a free service and would be just part of a Merriam-Webster website, and we really had a home page for the Merriam-Webster site before we even had a dictionary that necessarily would work with it. The initial registration of m-w.com was done by Britannica. In retrospect, I think we wish we had gotten webster.com, [as] dictionary.com, had the smarts to do, and so we were never able to get that URL, which I think is kind of interesting and ironic because we had put so much effort into protecting the Merriam-Webster name during the print era because of the confusion over the Webster's name, and "Not Just Webster, Merriam-Webster," and really driving home the Merriam side. Oddly, we get online, and our principal competitor is not making any use of the word Webster. So this battle we had done for years to make sure that you knew that Webster was Merriam-Webster kind of disappears online because no one really exploits the Webster name online, no major competitor to Merriam-Webster online uses Webster as part of their URL. Given all of that in '95, of seeing how the web is really flourishing, there are good interfaces for it, we've had a good experience with AOL, we know that there's pent-up demand to use the dictionary, it wasn't that tough, it wasn't that hard to make the decision: let's put it online. Interesting footnote to that, I think, Emily, at one point you were asking me about, "Well, when they launched Merriam-Webster online, what was the dictionary?" The funny thing is, strictly speaking, the site was never launched. We always said, as soon as we fix a few more details on this site, then we'll launch it. Well, it's not working quite well enough yet, we'll launch it next month, or we'll launch it the early part of the next fiscal year. Once we get this working a little better, then we'll really we'll promote it and we'll publicize it. We never did. This event...

Emily Brewster: It was live that whole time.

John Morse: ... it was live. We had opened it to the public, but we hadn't actually made any pronouncement or announcement or promotion that we had done it.

Emily Brewster: Interesting.

John Morse: By the time we get to early '97 and it's been up for six months, we realized it's launched. We were able to scare off the people who were running unauthorized copies of C7 [the Seventh Collegiate] and tell them that if they didn't take down their application, we were going to come after them, but then I think also subtly let them know that if they wanted to point to this hypertext Webster at the University of Michigan, there's nothing we can do to stop you. You can't run our dictionary, but we can't keep you from pointing to another dictionary. With that kind of logic, we were able to shut down almost every website that was running an unauthorized copy of the Seventh Collegiate, except we didn't do anything to the people at University of Michigan with their hypertext Webster, which meant that by 1997, all the traffic that wanted to use the unauthorized C7 was all going to University of Michigan, and so we sent our cease and desist order to the University of Michigan and said, "You got to take that down because that's our database. That was the Seventh Collegiate that's still protected by copyright. You need to take it down, and please redirect all that traffic to Merriam-Webster online," and at that point, our traffic, I think, tripled.

Peter Sokolowski: Wow.

Emily Brewster: Which dictionary was the merriam-webster.com dictionary at that point?

John Morse: I believe it was the 10th. We launched with the current edition in '96. You and I went back and forth about trying to understand that. I think the confusion comes because when the 11th came out in 2003, we didn't immediately update the website to the 11th, so there was a period after the 11th was published that the website still had the 10th, but I think the dictionary that launched in '96 was the 10th from the beginning. I don't think we ever tried to do a launch on the basis of the Ninth. When we were doing our initial development of that dictionary, it was shortly after the 10th came out. There was no reason why we would've fooled with the Ninth. So I think it's been the 10th all along. I certainly know this, we'd know from a review in one of the library journals that it was the current edition certainly by '97, and I don't see any evidence that we switched editions in the first year.

Peter Sokolowski: But your mentioning of the launch of the 11th in 2003 does bring up an important point, which is that whereas today we think of the updates as being online not just first, but immediately, and especially those updates were actually sometimes held back to allow for the new copyright of the print version to spend some time in the world and in particular to have those initial sales from publicity before we updated the website, whereas that sounds backwards today, but that was the way we were thinking as a book publisher.

John Morse: Right. What we wanted to make sure was that if we began publicizing an update that it was available both online and in print, that one didn't get ahead of the other in either case. I mean, you didn't want to publish a new edition or a new copyright update with a bunch of new words and not have them on the website. Then you would be considered way behind the times if you didn't have your website up to date. But similarly, we didn't want to disappoint a print user. If they saw that we had launched a bunch of words on the website, and then they went and then shelled out $24.95 for a dictionary and didn't get them... So we did try to time that with some care. I think we were being overly scrupulous at the time. I think one of the myths of dictionary publishing is that people buy dictionaries for the new words. I know marketing people think it's the case, and we build these whole launches around them, but I think "new words" is sort of a little bit like the prescriptive-descriptive debate. I think there are ideas that are talked about, and then there are ideas that are actually important. And they're just not always the same, and I think we talk about prescriptive-descriptive amongst ourselves as if somehow that explains dictionary publishing. I think for most of the buying public, they don't care. We talk about new words as if that's what's going to motivate people to go out and buy a dictionary. I don't think there's really any evidence that that's true. One of the great things, and I think this is sort of the next part of this story, and Peter you've been deeply involved in this, is that by going online, we now know what are the words that are most frequently looked up in the dictionary. And a number of people in marketing and elsewhere really were looking forward to this with a finger just ready to wag to say, now you editors, now you're going to find out what people really want to know. They don't need three definitions of some elegant word. They want to know trademarks, they want to know brand names, they want to know movie stars, they want to know brand new words, they want to know slang words, they want... And turns out no, what people want are in fact those slightly esoteric words, and yes, they would like more definitions, exactly what it is that lexicographers had been laboring on for the past 400 years...

Peter Sokolowski: 400 years.

John Morse: ... is what people want. I mean, it turns out that we did not have it wrong for these past 400 years, people did look up ubiquitous, paradigm, esoteric, epiphany, conundrum.

Peter Sokolowski: Integrity.

John Morse: Integrity. Those were clearly at the top of the list. Now, in the earliest days of the website, the four-letter words were pretty much at the top of the list. I guess there was still a novelty factor involved or the web traffic at that point probably skewed more towards younger kids who enjoyed looking up the word and giggling, but pretty much over the longer term, it really settles out that it is the slightly elegant, slightly literate...

Peter Sokolowski: Abstract.

John Morse: ... abstract words.

Emily Brewster: Latinate.

John Morse: Latinate.

Peter Sokolowski: Latin and Greek roots.

John Morse: And to an unanswerable degree, beginning with a vowel are the words that people really are looking out most frequently.

Peter Sokolowski: It's incredibly gratifying, and it does actually reflect the whole history of English language dictionaries because the first dictionaries of English were just of those words, the sort of Anglo-Saxon roots, the meat and potatoes words that are our function words like the verb to be, for example, or the pronouns, weren't even defined in those first-

Emily Brewster: Right, right. A simple word like house, you didn't need a definition of that.

Peter Sokolowski: Or dog, no, but they would have canine. So they would have that Latinate version. So there was a kind of home language that was always recognized as the non-academic language. So you could call them SAT words really. That's that kind of language.

John Morse: Right, they are SAT words. At the time we launched the website, it was about the time I became president and publisher. I think I was in a position to have the data behind me to say, no, we don't have to go chasing whatever is a fad. Yes, we're going to publicize new words, and I worked as hard as anyone to publicize new words, but we're not going to skew or alter or distort our publishing program. If I can find a way to get your new word story out of what we ordinarily do, I will do it, but we're not going to distort anything we do because in fact, look at the data. These are serious people who have serious questions about serious words. They don't need Macarena in the dictionary. Unless Macarena has an interesting etymology, maybe it does. That's not what we need. We need to make sure that if someone doesn't understand why talking about a sporty Volvo is an oxymoron, they've got to be able to come and see that we are now defining oxymoron in a way that goes beyond the original purely rhetorical sense and talks about the extended figurative sense of oxymoron. That's what has to be there. It helped me coming into that position of president-publisher from the editorial department to say, not only is the editorial department going to rule the roost around here for a while, not just because I'm president, but because the data supports it, the data says people want the kind of information that this department is creating. This department is creating the information that is going to fund this company going forward.

Emily Brewster: There is no compromise necessary on our defining criteria or any of our practices because the data verified that what we were producing was what people wanted.

John Morse: The other thing that I think really propelled us comes a few years later, and you cannot overstate the importance of Google to the program. When Google is founded in 1998, we had maybe about 8 million page views a month. I'm trying to remember the numbers. By 2000, at which point Google now is the most trafficked website, we've got 30 million. Google in two years, takes us from 8 million to 30 million page views a month, and one of the reasons, we find out, is that one of the most frequently used search terms on Google is "dictionary."

Peter Sokolowski: The word dictionary.

John Morse: Is the word dictionary. People are using Google to find a dictionary to look up a word.

Peter Sokolowski: And we know that actually from our own data too that a lot of people land on our homepage and key in the word dictionary looking for the dictionary, not knowing that they are already there.

Emily Brewster: That's right.

John Morse: We still don't have a good advertising model at that point to be sure. Merriam-Webster knows that it wants to be advertising-supported, and we knew that before the site even was working. We had already started having conversations with advertising agencies to see how we could have either advertising-supported or at least a sponsorship-supported site, but it really is Google that powers that growth. And it's because apart from the other kinds of search engines that existed in those days, which really wanted to keep users in kind of their online ecosystem, Google alone said, all we're in business for doing is finding a website that answers your question and sending you to it. We're not going to try to keep you on the Google website. And to this day, the Google page is blank. Maybe we'll get to this. The situation clearly has changed with Google, but at least as they start off, all they're trying to do is get you to a website, and that's what we were doing. Alone amongst dictionary publishers, Merriam-Webster goes into this period only publishing dictionaries. If you were Random House publishing a dictionary, you also have a big print business. Simon & Schuster, you've got a big print business. Houghton Mifflin, a big... Merriam-Webster alone is just purely in the dictionary business. We've got no other business to protect, so I think that's why Merriam-Webster was able to embrace the notion [that] we will have a freestanding website that offers dictionary content free on the web in the hopes that a business model will emerge in time.

Peter Sokolowski: It's kind of amazing.

John Morse: But what Google does is it really pours traffic into sites like Merriam-Webster and Dictionary.com; to be sure, Dictionary.com even more than Merriam-Webster because it's sort of the crude early days of search engine technology. If your search term matched the name of the website, that's the highest relevance, and if you wanted to look up information about Abraham Lincoln, and there was a site called abrahamlincoln.org, that's where you're going to go. So if your search term is dictionary, Dictionary.com is going to show up first. So they went from being a kind of inconsequential technology platform to being a potent presence on the web no matter what dictionary they were working with. Now, after that influx of traffic, they did a deal with Houghton Mifflin and were running the American Heritage Dictionary, and then ultimately of course, acquire rights to the Random House Dictionary when Random House discontinues its print program.

Emily Brewster: And those were both good dictionaries, they were solid dictionaries.

John Morse: Excellent dictionaries. A couple of points to draw out there, the practical way of the people who were using dictionaries shifts pretty soon when people realize that they can Google the word they want rather than Google the word "dictionary" and look up the word they want, and with that, some of Dictionary.com's advantage over Merriam-Webster decreases, and now it really becomes a fight between your search engine optimization experts. And in large ways, that's what's going to determine what dictionary people will use, is what comes up at the top of the list. The sad fact is that artificial intelligence really begins with Google where we all just sort of outsource our discernment to Google and say, Google, you tell us what's the best website, but what Google is doing to do that is a series of algorithms to evaluate relevance of a website, which may or may not be quite right. Whatever it is, that difference of people googling the word dictionary to googling the word they want to look up worked very much to Merriam-Webster's advantage, and we began taking over the first spot and holding it. Unless still would happen that it would not take much to change in a Google Search algorithm to move that around. So at one point, I think Google was particularly interested in websites that had some kind of graphic above the fold. And by above the fold we mean they would show on your screen without having to scroll down to find it. So Dictionary.com, for every entry, showed a little map of the world, and then they colored in the country where that word would've originated as far back as their etymology goes. An awful lot of them just lit up Italy, Greece, or Germany, but because they had that feature that essentially was putting a graphic above the fold for at least a period of time, their Google rankings went up.

Peter Sokolowski: It's interesting how sort of arbitrary that is. It has to do with business, it has to do with algorithms. It doesn't have to do with lexicography.

Emily Brewster: But it was taking advantage of the digital platform, which was a key part of it. We were talking before about the database and the tags of these entries, but it took a while for dictionaries to recognize what else the digital platform could do. We very soon realized that we now knew what people were looking up. It took a while for our defining to really reflect that. I remember I had a project where I was to revise the 10 most commonly looked up words, and for the first time, I could make a definition however long I wanted it to be, could add however many example sentences seemed useful, and now that's a major part of our lexicography, is creating entries that are as robust as will benefit the reader. No longer are we required to be concerned about concision, about making a dictionary that is portable and affordable, but I think it took a while for us to fully embrace that freedom. For a long time...

Peter Sokolowski: [inaudible 00:22:17].

Emily Brewster: ... the online dictionary was matched almost to the print dictionary, abbreviations would be expanded, that sort of thing, but the actual text of the definition was identical.

John Morse: Yes. In large ways, we have yet to rebuild the dictionary the way it ought to be if you're going to use it online. Early criticism of online dictionaries, all of them is all they really do is give you access to the text of the print edition. And in some ways that means it's a less good experience because the digital display shows entries in isolation. You don't see the entry before it, you don't see the entry after it or one or two above or one or two below, whereas those entries may in fact have very relevant information to your search. The original dictionaries were meant to be print products. The unit of consumption was the page, so we kind of assumed if that information was on the page someplace, you might very well see it. We had rules of cross-reference. It was built around that depending on the dictionary that if the word shows up 10 entries above or 10 entries below, you don't have to create the cross-reference because people will see it. But then just more subtly, the etymology is only going to be at one form of the word. It's not going to reappear at the cognates. I don't think any digital dictionary has yet really done the rigorous rewriting that you would really need to do if you're going to insist on this experience as we all do that you view the entries in isolation.

Peter Sokolowski: Absolutely.

John Morse: You're right, I don't think we have yet actually lived up to the potential of what the digital-

Emily Brewster: We're closer than we were though.

Peter Sokolowski: Of course.

Emily Brewster: We're much closer.

Peter Sokolowski: It's a knot that will take a long time to untie. But I mean, just a basic thing, as you've mentioned, etymologies on the page, you have a word like literacy and literary and literal, they all connect to the same root word, and we're going to see that just an inch and a half away, and so that's not something that is repeated at all the entries, and that's, of course, a bad digital experience.

Emily Brewster: Although we do have hyperlinks, it's very easy to get from one to another.

Peter Sokolowski: Of course.

John Morse: Right. And I think that would be the right answer, is just a much richer set of hyperlinks. All dictionaries could be doing a much better job of building in those hyperlinks to essentially allow the digital user to know what the print user knew simply by viewing the page.

Peter Sokolowski: There is something that can be an online dictionary feature that is not possible in print, which is the audio pronunciations, and I would like you to comment a little bit on the origins of those, which are an incredible richness for this online experience.

John Morse: It was one of the first things we knew after we launched the website in '96 without audio, that what we needed to do was get audio. There's one thing that people are really dissatisfied with in a print dictionary: pronunciation symbols.

Peter Sokolowski: Reading those phonetic transcriptions are difficult. Most people have trouble with them.

John Morse: And honestly people who feel otherwise, but having put it in International Phonetic Alphabet would've made it worse.

Peter Sokolowski: Of course.

John Morse: International Phonetic Alphabet is worse than what we do now, especially for native speakers of English.

Emily Brewster: Right, for anyone who's not a linguist.

John Morse: Right. But people just really don't like phonetic symbols, so we moved as quickly as we could to get audio onto the site. We thought at one point, well, maybe that could be the basis of a premium site. You could get people to pay money for it, but that got shot down fairly soon. Our initial creation of a set of audio pronunciations actually goes back, again, to that era of licensing in the 1980s. We worked with a web developer called Highlighted Data who created the first CD-ROM version of a Merriam-Webster dictionary in the 1980s, and a feature of that was audio pronunciations. And that developer had hired someone to come in and pronounce every word in the Merriam-Webster dictionary. Apparently that job never got completely done, and I believe Julie Collier, who was our illustrations editor, for some reason, was called upon to finish doing the audio pronunciations for that product. So we knew something about how to create that. Those files turned out to be unusable unfortunately, so we knew what we had to do was create them again, and it was done with a team of, I believe, six pronouncers, half men and half women so that you didn't have one gender or another to create the audio pronunciations. I forget what year they went on. It took a while to get that project done.

Peter Sokolowski: And those were recorded in a studio setting, right?

John Morse: Yes.

Peter Sokolowski: I just remember hearing they were able to read the phonetic transcriptions, of course. So these were people who might've been trained in the theater or in opera because they were familiar with foreign words.

John Morse: They certainly were in the performing arts. I can't tell you what form, but yes, they were actors and actresses that were doing this. And I forget how long it took us to get that done, but we knew almost immediately that that was something that if you want to exploit the virtues of a digital dictionary, that's the first thing you want to go after.

Peter Sokolowski: And you're talking about something like 100,000 pronunciations for that first batch.

John Morse: Yes. All we did was head words.

Peter Sokolowski: Yes, it wasn't all the inflections, right? Or the variations necessarily. Some of them have a couple that are pronounced, but most don't.

John Morse: Yes, I think some had two pronunciations. The interesting part of that was that there was some problem about doing the offensive words that we got some pushback that somehow they felt bad about it or didn't feel comfortable doing it, and maybe we listened to them and didn't think they came out very well, so most of the offensive words were recorded by our pronunciation editor.

Peter Sokolowski: And what a great voice and pronunciation that he has, Brian Sietsema. And did he also oversee the project in the sense, did he check for the quality of those recordings?

John Morse: Yes.

Peter Sokolowski: Because my understanding, for example, if a given five syllable word was not a very clear, or [was] perhaps improper, maybe the stress was on the wrong syllable or something, rather than send it back through the whole process, I believe he just redid it himself.

John Morse: He may well.

Peter Sokolowski: So there was a kind of patchwork that was not considerable, but a number of dozens or couple hundreds of these he would just redo because he said, "What's the point of going through this whole thing again? I'll just do it myself." I saw a use of our audio phonetics in Cairo, Egypt, at the National Spelling Bee for English language learners, and I emphasized in my training for the pronouncers that it was okay, you don't need to be a native speaker of American English to do this as long as you're clear, as long as you're consistent. But there were several of the [spelling] bees that were conducted by playing the Merriam-Webster audio and repeating it, playing it several times for the young spellers, and I'd never seen that before, and it's great, but also it's important to see a pronouncer's mouth in a spelling bee. If you want to distinguish between a B and a V or an M and an N or a P and a B, there are lots of things that are missing. But at the same time I thought, well, you know what, it works pretty well for this purpose also.

John Morse: In fact, at one point, I think there was a website that was scraping our pronunciations and putting it on their own website, and that's all they took was the pronunciations. If you wanted to hear a word pronounced, you went to this website.

Peter Sokolowski: There's also a website that set them to music. Do you recall this? So the Girl from Ipanema, they would take every single one of those words and play it to some kind of sort of synthesized version of the melody, and the Merriam-Webster pronunciation would come out kind of more or less in rhythm. And it was always funny, everybody loved it.

John Morse: The problem with the professional pronouncers is that they couldn't get over the fact that they felt self-conscious when they were saying that...

Peter Sokolowski: But of course.

John Morse: ... and so you could feel sort of the doubt in their voice, whereas I think he could take such a dispassionate view of it.

Peter Sokolowski: That's an important point to make about the job, the habits that you get, and to be honest, the humility. The famous quote that you have made reference to many times from Fred Mish, which is that there's nothing so humbling as a walk to the citation file. And the fact is this is a little bit off the subject, but you have encountered this, I'm sure every bit as much as I have, which is the assumption of a stranger that I'm going to correct their English, which is actually the last thing that would ever cross my mind, mostly because my habits are on research, and I'm not cataloging your errors, I'm not silently correcting your grammar, and that someone like Brian who is professionally used to reading these transcriptions and to studying them and to recognizing they might be two, there might be three correct ways of pronouncing this word because he works as a spelling bee pronouncer for the National Spelling Bee, and I often pronounce for spelling bees. I know that once in a while, I might pronounce a word with the wrong vowel value or the wrong stress pattern, and I just toss it out. I just look, that's my mistake, that's my kind of relationship to this transcription and has nothing to do with this competition or you as a speller, let's go to the next word. In other words, I have no ego about it, and I think that's an important point that when you're steeped in the research this much, it doesn't become an ego problem, or to this point, the emotional distance from these units, these lexical units is greater for someone like Brian than for someone who doesn't work in publishing or lexicography and they can't read these words in a dispassionate way. This is a long anecdote for this, but I think it's an interesting point that the dictionary had to think about this.

John Morse: Those words are difficult for dictionary editors to know how to deal with, and clearly we know how difficult the decision around getting the F-word into Merriam-Webster was that the editors of Webster's Third thought it was high time that that was there. The definition was written, but the president of the company just said that was a bridge too far at that point, and when you think about all the controversies that came up around Webster's Third, adding that to the mix probably would not have made life any easier for anyone, so maybe it was not such a terrible decision.

Peter Sokolowski: These are cultural considerations and marketplace considerations that they have to do with publishing in general. But clearly his fear, I assume, was that this would be the headline, that this would be the shorthand for the new edition. Nobody wants that if you're making a big investment in this product.

John Morse: Right. Whether that was his fear or not, I think it would've just added to the bill of grievances that the critics of Webster's Third would've had.

Peter Sokolowski: And so it's ultimately added about a decade later.

John Morse: It's ultimately added in... I think it gets into an addenda in '72.

Peter Sokolowski: Okay, so just before the Eighth Collegiate.

John Morse: And I think the pride of place, if you want to, for entering it, I think goes to the first Houghton Mifflin dictionary. [inaudible 00:33:23].

Peter Sokolowski: Yeah, the American Heritage.

John Morse: I think that came out in '68, but...

Peter Sokolowski: Yeah, I think that's right.

John Morse: ... I may be off on that.

Peter Sokolowski: But we were talking about the online dictionary, and it's not only its constituent parts, but it's function as a business model, and how did that evolve?

John Morse: Well, we didn't know, is the plain answer to that. Again, if you go back to the decision to go online and knowing that a lot of people want to use the dictionary online for free, do you meet that need or do you just stand aloof? And I think we did, and I think we said we did need to meet that need, and I think we did so in the faith, naive maybe, that a business model would emerge. It was the gamble that we took, and the gamble, as you know, we've talked about this, was, are dictionaries actually heavily used by people or aren't they? If dictionaries aren't heavily used, if dictionaries really are just bought, brought home, put on a shelf and not used, we could really be shooting ourselves in the foot because people would know that the dictionary was online so they wouldn't buy the dictionary, but if they don't really use the dictionary, they wouldn't come to the website either. We were making the gamble that in fact, people do use their dictionary intensively and that all we were doing is moving people from buying a dictionary to using the dictionary online, and that eventually a business model would emerge. We didn't know what it was really at the time. Yes, we knew we wanted it to be an advertising-supported site. We knew that we wanted to be free [inaudible 00:35:03]. The earliest references in-house, we call it the Public Access Dictionary, but we really didn't know how we were going to make that work. Amongst dictionary publishers, we alone made that gamble. None of the other web dictionary publishers wanted to. And famously, the senior vice president at Houghton Mifflin, Wendy Strothman, who was responsible for dictionaries over there, really kind of scoffed at the whole idea and said, "We are perfectly willing to have our dictionary data used online, but we want to be paid for it." And of course, they were, at that point, I think being paid for it by Dictionary.com, and she said, "What puzzles me is dictionary publishers who think they can just give it away." I think that was pointed particularly at Merriam-Webster, how can you possibly do this when you're just giving it away? If that gamble was right, if people really do use their dictionaries a lot, then a business model has got to emerge. Business models will rush in to fill a vacuum.

Peter Sokolowski: And you had confidence because of the AOL experience that this would represent a constituency.

John Morse: That's how we knew that there would be people coming to the site between the AOL and the experience with unauthorized Seventh Collegiate, but that wasn't necessarily telling us that we could make any money out of it. The fact is, for at least the first five years of the site being up, we didn't really know where the money was going to come from, and it wasn't coming in. We were not really making much money off of that website for the first several years that it was up, and we were getting a lot of heat from that. I mean, there were people internally saying to Jim Withgott, who was a vice president at Merriam-Webster, who really championed the website, and probably more than anyone inside the editorial department really advocated for the success of the site. But both Jim and I were pretty much sort of saying, well, give us time, something will emerge. But we couldn't answer the question, where's the revenue or what the business model really would be. And we were under pressure certainly at some points from our parent company to move it all to being a subscription site. That they understood.

Peter Sokolowski: Because you could understand that, right?

John Morse: How are you going to do it if it's just a free site?

Peter Sokolowski: And they had some experience with, for example, university subscription library subscription models, which is a good way to do business.

John Morse: Right. And that's the way certainly OED managed...

Peter Sokolowski: Of course.

John Morse: ... as a subscription site. OED specifically didn't want individual subscribers. They only wanted institutional subscribers. They set their institutional individual subscriber rate prohibitively high, and according to some people I talked to at Oxford, on purpose because they just didn't want what they thought would be the hassle of maintaining individual subscribers. They didn't want people, at that point, I think they imagined calling in and saying, "I forgot my password," or, "I'm sorry, my check bounced," things like that. I think that fear was misguided with the experience we got from running the Unabridged as a subscription site indicates that it was actually frictionless, that's not a difficult thing, but nevertheless, I think for us from the beginning, we wanted it to be an advertising-supported site. The problem that we had was not just necessarily traffic, because as I say, by the time Google comes along, they boost our traffic significantly. The problem is that advertisers want to reach a particular kind of user. When we talked about this a little before, if you're manufacturing baseball gloves, you want to advertise to people who are likely to buy baseball gloves. And across the board, advertisers are looking for niche constituencies, and dictionaries can't do that because when you index the profile of dictionary users, it's just average. We don't have particularly more rich people or particularly more less affluent people, we don't have particularly more older people or younger people. I mean, it's a little bit older, it's a little bit better educated, but not hugely. And certainly it isn't tied to people who do fly-fishing or need a new set of snow tires or anything like that. So we couldn't get the kind of high value advertising that you would look for. We ended up getting kind of the dregs. We got credit card companies or weight loss programs or home mortgages, the kind of advertiser who just wants to get a whole lot of people irrespective of any particular demographic qualities to them, and that's the least valuable advertising that you can run. So we were running lots and lots of advertisements, but at very small, what they call CPM. The amount of dollars we got per thousand ads we ran was pretty low. What saved us is what spooks some people, which is that the software was being developed, and again, a lot of it by Google also, to do behavioral targeting that because websites put cookies on your computer when you visit them, and then other people can read those cookies. It's relatively easy for ad-serving platforms to know what websites a user has seen before coming to your particular website. So when a user came to Merriam-Webster, we knew, not we, but the ad-serving companies knew what were the last five websites you looked at before that. And on the basis of that, they can pretty much figure out your age, your address, your gender, your income level, and now they could begin to target ads to the user. And what would happen is what they called online bidding. In the nanosecond between when a person clicked on the link to come to Merriam-Webster and when the page emerged, an auction would take place in which the ad-serving platform would say, I've got a user coming to Merriam-Webster who displays these five characteristics, who wants to bid on this user? And the advertisers would bid on that user, and whoever bid the highest got to run the ad.

Peter Sokolowski: So that gave better revenue.

John Morse: At that point, our revenue began increasing significantly such that revenue from the website was replacing the lost revenue we were seeing from decreasing web sales, from print sales.

Peter Sokolowski: Print sales.

John Morse: Which really didn't happen to a significant degree. It began increasing a little bit around mid-2000s, 2005 or so. But it really was the recession following the 2008 financial crisis. It was with that 2008 to 2010 is when we saw the most of the fall off in print revenues. And that was the same point at which the ad revenue was really finally kicking in.

Peter Sokolowski: And they sort of canceled each other?

John Morse: They sort of canceled each other.

Peter Sokolowski: It's kind of astounding.

John Morse: Well, the thing to bear in mind, people always talked about the fallacy of digital publishing as you were trading print dollars for digital pennies. But what that ignored was the tremendous price of distribution that all print publishers or particularly dictionary publishers had to pay. So that $24.95 dictionary in the bookstore, and I'm dating myself, but at some point in our story, that's a $24.95 dictionary, the amount of money that's coming back to Merriam-Webster is a fraction of that. So what I've been known to say is in the dictionary world, we were not swapping print dollars for digital pennies. We were probably swapping print dimes for digital pennies.

Peter Sokolowski: That's a very different ratio.

John Morse: Dictionary publishers had [a] much higher percentage of their revenue go to printing and binding the dictionary. We were selling print dictionaries for about the same price as trade fiction or nonfiction, which had about maybe 400 pages, and we had 1,600 pages. So right away, our print bill was going to be four times the size, but our revenue was going to be the same. And then that same period when digital publishing is beginning to take off, our distribution partners were getting greedier and greedier about how much of that sticker price they wanted to keep for themselves. So if you go back to the 1980s, you might've sold a dictionary at a 40% discount to your retailer. Now it's probably more like 50%. Half of that $24.95 stays with the retailer. And then a good portion of the $12.95 that's coming back to you is going to the printer and publisher.

Peter Sokolowski: And shipper.

John Morse: And shipper.

Peter Sokolowski: Because dictionaries are heavy.

John Morse: It was not that hard for the digital revenue to replace a profit level. So our top line revenue may have been lower, but our bottom line profits were able to sustain it because the profits...

Peter Sokolowski: It's amazing.

John Morse: ... from the digital publishing could replace the lost profits from declining print sales.

Peter Sokolowski: And it's partly because by 2010, the internet was very mature as a kind of marketplace, and people were really there for work and there for shopping in ways that they weren't 10 years earlier.

John Morse: Yes, that was probably true for dictionary websites in that dictionaries were more of an office productivity tool than other competing websites.

Peter Sokolowski: Of course.

John Morse: So it was a product that people were encountering in the workplace, but I also think people were becoming more accepting of advertising in two ways. If it was there and they noticed it, they understood why. And I think also oddly, they kind of developed a sense of not seeing the advertising. Just fairly recently at an event where you and I were attending, in a question and answer session in the aisles afterwards, someone asked me, "I don't understand how you make any money off of your website. I don't see any advertising." And of course, the website is filled with advertising.

Peter Sokolowski: Sure, sure. That's an amazing comment. It's a tribute to also good page design so that they're not distracting, and also has to be said, decisions, hard decisions, but good decisions to not overpopulate and mess up the page.

John Morse: And I think that took years to kind of triangulate.

Peter Sokolowski: Because there's some X factor that connotes quality.

John Morse: And I have to give really credit there to the person who's now president of Merriam-Webster, but who had originally come to Merriam-Webster to run our advertising sales operation, for really understanding how to balance those imperatives and be very revenue-conscious because that's what's going to keep business, but also aware of how you can overdo it. I think he really navigated a good course for us, understanding just what you can and can't get away with.

Peter Sokolowski: His name is Greg Barlow. And to keep the basic idea of a quality experience online to connect with a quality brand and a quality dictionary.

John Morse: And again, that goes back to, I think, that understanding that branding is what keeps us in business, branding is the reason why we have a Merriam-Webster dictionary today, and it was understood as far back as Noah Webster, and I think we maybe mentioned before, being probably the first person to do an author tour, going up and down the East Coast, selling not only his blue-back speller, but I think also essentially promoting his own expertise as a creator of content. People began to really yoke the name Webster with the name of language information. So I think it's been crucial to us, and I think, luckily, we have been led by a series of leaders who get that it's all about the brand.

Peter Sokolowski: And a long tradition.

Emily Brewster: For the Word of the Day and all your general dictionary needs, visit merriam-webster.com. Our theme music is by Tobias Voigt. Many thanks to John Morse, and to Ed Finnegan of the Dictionary Society of North America for suggesting this oral history project, and to New England Public Media for the use of their studios, and especially to NEPM's John Voci for producing these three conversations. I'm Emily Brewster for Peter Sokolowski. Thank you for listening.