Where Grapeshot came from
You will no doubt be delighted to be told that although this is a talk about the history of the core Grapeshot software, it is not at all technical. The people from outside Cambridge will be relieved. I am sure that the only thing more agonising than the technicalities of advertising to computing people must be the technicalities of computing to advertising people.
And it is also about me and John working together. Andrew Goyder was called into one of the board meetings last year, and apropos of I don’t know what, said, “Of course, John and Martin are very different.” I protested, “No, John and I are actually very similar.” This was ignored. I did not even get those looks around the board table that seemed to be saying, “Couldn’t Porter be put back on his medication?” Well, I’ll try and show those similarities now.
I dare say many of you here see Grapeshot as a startup company that began life in 2009, with our first investment boost from IQ Capital. But to understand it fully we have to go back a little further in time than that, namely by about 1,200 or 1,300 years.
Cambridge! This is where it all began.
— And it was around then that the Anglo-Saxons must have been aware of a declivity in the water level along the stretch of the Cam here, which made them realise it was a good spot to build a weir, that could be used to drive a water mill. The river is the Cam, the weir had its bridge, and hence Cam-bridge. That was how it all started. Farmers came from miles around to have their corn ground, and the spot became important. South of the weir (upriver) the land would be boggy, like fendland, and liable to flooding. It is still called Coe Fen and does flood sometimes. North of the weir (downriver) the water level would be lower, controllable, and the land would tend to dry out. A good place for building collegiate structures that could use the river’s resources, and so the backbone of colleges from Queens to Magdalen.
Imagine you’ve gone to the end of the bridge and turned right to look at the KEEP CLEAR sign. This is what you see,
There’s The Mill pub, and the Graduate Centre. Can you see, among the cobbles, a circle of stone set in the ground? Go up to it, and it looks like this,
(There used to be two of these, but in place of the second is now a pool of tarmac. No doubt an enquiry to Cambridge City Council would explain the mystery.) It is an original mill wheel. Not of course from Anglo Saxon times, but in use when the two mills finally stopped working in 1927 and were pulled down. So if anyone asks you about the beginnings of Grapeshot take them to this mill wheel. Without a mill wheel there’d have been no mill, without the mill there’d have been no colleges, without the colleges there’d have been no University, and without the University there’d have been no Grapeshot.
St John’s College, which I went to in 1963.
The evangelist St John my patron was
Three Gothic Courts are his, and in the first
Was my abiding place, a nook obscure.
That of course is not me, who never had room in first court, it is Wordsworth, writing in the Prelude. You may recall the company meeting of just three years ago in the Old Music Room. That was here. Wordsworth was on C staircase, almost diagonally opposite, just here, where he could hear all the kitchen noises, “less tunable than bees / but hardly less undustrious”, and the sound of Trinity great clock through the night. Now the next staircase along is D staircase, and here, in 1984, were the rooms of a young lady called Sarah Brewster, studying anthropology. Sarah went on to fall in love with another contemporary student, John Snyder, who switched from Geography to Anthropology after one year. They married, and Wordsworth might have been pleased to know that they later moved to his beloved Lake District.
While I was at St John’s John went to Christ’s College — not of course at the same date, but 20 years later than me. Here it is. Now you’re probably thinking, they all look the same, how to you tell these colleges apart?
Well, a good differentiator is to look at the college arms. Here are the arms of St John’s,
— and here the arms of Christ’s. See the difference? Christ’s has more of these little doohicky things. (Incidentally “doohicky things” is not the correct heraldic term.) OK? St John’s — Christ’s — St John’s — Christ’s. Or have I got it the wrong way round ... well it doesn’t matter too much, you see the difference there.
Cambridge in the early 60s, when I was there, was a world away from the Cambridge of 20 years later which John and Sarah knew. I must resist the temptation to illustrate the differences with endless anecdotes, or I’ll be talking forever. My Cambridge was Hogwarts, but without the magic, and all the colleges strictly single sex. One could no more imagine a Sarah Snyder with rooms in St John’s, than a unicorn or phoenix living there. It was more like the Cambridge of the 1760s than the 1980s. So to understand it I recommend this book, Winstanley’s Cambridge in the 18th century,
which is not just a delightful book on Cambridge in the 18th century, but a delightful book about the 18th century itself. By a happy chance I found a copy, in pristine condition, in a local bookshop some months ago. I had hoped it might be given out as a prize sometime today, perhaps to the victor ludorum (or vitrix ludorum) of the ensuing games. But I reread it very recently, and came to the conclusion that while it is delightful to me, I think most Grapeshot employees would find reading it a severe punishment. I will simply leave it Gael to find a suitable form of disposal [hands book to Gael].
Of course, not all of Cambridge, even then, was in the past. I imagine no one will recognise this late Victorian worthy, but admire at least the luxuriant late Victorian beard:
It is Henry Sidgwick, the philosopher, almost forgotten now, but important in his day, and something of a Cambridge hero for the stand he took against swearing to the 39 articles.
To find out more on late Victorian Cambridge, you might try this book,
Sidgwick Avenue is named after him,
and in Sidgwick Avenue is Newnham College, for women, and built without a chapel, as well as the ultra-modern buildings of the Sidgwick Avenue site,
Here for example is the amazing History Building, or Seeley Library as I think it is properly called,
And here is the Raised faculty Building.
From its top floor you get this view of the internal courtyard, rather like a college couryard,
and round the corner you can see the narrow internal corridor.
And it was behind this door, closest on the right, that I worked for three years in the so called Literary and Linguistic Computing Centre, part of Modern and Medieval Languages, which was finally closed down 6 years ago in 2009. This nook obscure was my abiding place. I got there a solid experience of non-numerical programming (as it was called then). Much of the work was creating indexes and concordances of ancient or medieval texts for scholars. It was gratifying to get it right, but the work became very mechanical after three years, and in 1974, I moved to a much more ambitious project.
Now does anyone know this Victorian worthy? Here he is again. There are clues here,
It is Adam Sedgwick, teacher of Darwin, founder of the Sedgwick Museum. Now it is very important, of course, not to confuse Adam Sedwick with Henry Sidgwick. Very different people. But here is something to surprise you all: they were in fact distantly related, and came from the same big Yorkshire family that spelled its surname either “Sedgwick” or “Sidgwick” as the fancy took them. A lot of people don’t know that. No beard, characteristic of the early Victorian period. For more on Cambridge at that time, this book can be consulted. (The more alert audience members may have already guessed the title.)
The project was to computerise the Sedgwick Museum catalogue in the department of geology.
If Cambridge was like Hogwarts, the geology department was Gormenghast. (Has anybody read Mervyn Peake?) It is impossible to imagine what the place was like. If you said “how many people work at Grapeshot?” the answer would be, about 60. If you’d said “how many people worked at the geology department?” the answer would have been, rather less than half of them. At one time I worked here,
the so-called curator’s pew, which used to house the catalogue, describing half a million specimens, in mighty ring-clip binders, with a card index in the centre. Now it is a sort of information point. It was perishing cold in there. There were regular trips to Spitsbergen, and I think the department was kept at a temperature to inure the students to rigours of Arctic existence.
Finally the computerization project was done. It started in 1974, and we demonstrated a working system in 1982. Now I was not working solidly on this all that time. I had other jobs in between, the really important one being a period doing research on IR with these two guys,
This is Cornelis Joost van Rijsbergen, who, as you might guess from the name and cut of the beard, is of Dutch origin, although he was born in South Africa and grew up in Australia. He has always been called “Keith”.
And this is Stephen Robertson, whose beard is the pure academic whopper. At the time, and this is rather characteristic of me, I had no idea of the importance of the work we were doing, or of the very high standing of these two men. Keith later became professor at Dublin, then Glasgow, and took IR ideas in a new direction that greatly influenced work in America. Stephen ended up in Cambridge at Microsoft Research, by which time he was regarded as the world’s leading IR theoretician. All I had for a long time was the feeling that Andy Morley confessed to me when he first joined Grapeshot. “What are these guys talking about?” But like Andy, I did pick it up. By diligent reading, judicious questioning, and a process of intellectual osmosis I learnt Information Retrieval, well enough at least to be able to make a career out of it. So when I went back to work at the Geology Department (now just a section in the Department of Earth Sciences) my mission was clear. All the surviving software for the Sedgwick Museum project was unceremoniously dumped, and I wrote a new, really big system, in a very short space of time — just a year or so. An IR system for museum catalogues.
But I could not think what to call it. And the perfect name was given to me by this lady,
Maggie Carr, of the Cambridge Computing Service, who told me to call it Muscat, from “mus” for museum, and “cat” for cataloguing. All she asked for in return was a bottle of Muscat wine, which I duly gave her.
It was a good name, because of its various associations, muscat wine, muscat grapes, Muscat the principality in the middle east, and other mysterious things (“Muskat”).
In 1984 I finally left the University for good, taking with me the marketing rights to Muscat. For the next seven years I had more adventures than Sinbad, though none so exciting. I worked freelance, but by 1991 I was running out of customers. I had connections with the BBC, as I had worked on their BBC Domesday project, when they had published a snapshot picture of Britain on two laservision videodiscs. (A videodisc was the precursor of the modern CD or DVD, with a data capacity somewhere between the two. It is difficult to compare exactly, as the videodisc held a mixture of analog and digital data. It was about the size of an old LP, and needed its own player.) The Beeb continued their interest in “interactive television”, as they called it, for some years after the Domesday discs appeared, and they were looking for things to publish. And here a name kept coming up, that of John Snyder, who was on a field trip in North Africa making a photographic record of the Tuareg people, nomads of the desert. I was also doing work for professor Alan Macfarlane at Cambridge, who was putting together a data repository of information about the Nagas, a primitive hunter-gatherer peoples squeezed into the border area between India, Burma and China, possibly to be published on videodisc. John Snyder was also Cambridge based, so I kept hearing his name here too. Let me show you some Nagas.
Here is a Naga girl. Her teeth have not gone bad: Naga girls used to blacken them. They thought it enhanced their beauty. You can see she’s in a line dancing, holding the hands of girls on either side.
Here are a line of Naga men, similarly engaged.
Now the Nagas were a headhunting people. (Nowadays the practice is, we hope, discontinued). They divided up into different clans and the men engaged in a kind of continuous ritual warfare with each other. When you killed an enemy, you chopped off his or her head, and brought it back as a trophy, where it would be proudly displayed in the village in a head rack with other heads. The more heads a warrior chopped off the more prestige he acquired, and the more wives he got. If you asked them why they did this, they would say that, in the next world, the souls of the slain became the slaves of the man who killed them in this world. It was like making voluntary pension contributions for your condition on the far side of the grave.
I became curiously fascinated by these strange people. The violence of their behaviour belied the attractiveness of the materially rich culture they created. Anthropologists found them friendly. The word “headhunter” always makes me shudder when I hear it in board meetings, which is often enough. And yet in a literal sense, perhaps a Naga headhunter, like this one,
might have some use in our company in our competition with business rivals,
Well, I was in Macfarlane’s room one day in 1991, alone, doing some work on the PC there, when someone entered and asked if the professor was around. When told no, he introduced himself as John Snyder. My astonishment cannot easily be imagined. My mental image was of an aging Cambridge anthropologist with weatherbeaten face and grizzled beard. Think of Henry Sidgwick in tropical kit, kakhi shorts, pith helmet etc. John had no beard at all! Here is John as he was then,
(a little later actually, as the girls were not yet born). John now, I think, looks very much like John then. So much so, that I still think of him as a man in his early twenties. By transference I think of those working for him, Chris Good and the rest, as children or teenagers, and so sometimes call them “the kids”, no doubt to their irritation.
Please note the Snyder eyebrows, a dominant genetic characteristic of the Snyders, as distinctive a family mark as the Hapsburg jaw of the emperor Charles V and his numerous descendants.
I would very much like to have shown you some images of the Tuareg from John’s collection (they are wonderful photographs), or of the famous landrover of the Sahara expedition. But I’m afraid they have not been forthcoming, despite approaches to no less than three members of the Snyder family. (sigh). You just can’t rely on these Snyders!
What at once impressed me about John was that he was actually interested in what I was doing, and asked many questions. That may not seem remarkable, but I had been working for the BBC on and off for six years, where talking was so very much more important than listening. We were connected by having handled real collections of data, me with 440,000 fossil objects, John with 50,000 photographs, and that is not a bad approach to getting interested in IR. This goes down to a deeper similarity, that we are both natural hoarders. Of course there is a difference of scale: I can’t bear to get rid of an old theatre programme, John can’t bear to get rid of the landrover he crossed Africa in. But the principle is the same.
It was later that year, October 1991, when John rang me suggesting we might combine forces. We met, I think, four times that month. John had on a little badge, “proud father”, Alexandra having been born in August. It must have been on our second Norwich meeting, 25 October, that we decided to go into business together, and Muscat Ltd got formed.
At that time we were both so poor neither dare admit to the other how little money he had. Another similarity, that we’ve tended to be rich or poor at the same time. The difference being that however poor John is he behaves like a prince, while however rich I am I can never quite shake off a sense of poverty, so I’m the one who stands in the pouring rain for half an hour waiting for a bus because I think it would be extravagant to take a taxi.
The business grew slowly. Our first success was with Reuters, who gave us some work. Then, I met John Orosa, the person I’ve known longest after John Snyder, among those here at Grapeshot. We acquired premises in the St John’s Innovation Centre — no picture needed. Nearby was another little company. Its boss was a certain Mike Lynch, who took an enormous interest in our software. We almost went into business with him, but not quite. (I wonder what would have happened if we had?)
Ah yes, and this is me when we first met. Note the functional beard of the software writer (no time to shave).
John confused me by forever changing houses. First he was living here,
a bijou terrace in Tenison Road (I may not have that quite right). But then he downsized,
then he upsized,
then he downsized,
(Ah no, this is where my daughter lives in Brixton, thanks to the London housing crisis.) But I realised finally that it was all governed by a simple formula, the size of John’s house being in inverse ratio to the extent of his financial embarrassments.
Finally — and this is jumping ahead to today — he took his family up to the Lake District,
for the full Worsworthian experience. There he is like Wordsworth’s “boy”,
There was a boy (or CEO) you knew him well ye cliffs
And islands of Winander. Many a time
At evening when the stars had just begun
To move along the edges of the hills,
Rising and setting, would he stand alone
Beneath the trees, or by the glimmering lake
To paraphrase, holding his hands like this — to make a hooting sound — can’t do it — he “makes mimic hootings to the silent owls / that they might answer him”. Well I assume that’s what John does in the Lake District, because I’ve read Wordsworth, and there’s not much else to do. Nobody really lives there, although you might find a leech gatherer on the lonely moor, plying his uncertain trade, or a dog, howling by its dead master, who’s fallen off a cliff. But I suppose such experiences give John the necessary resolution and independence to be our CEO, and the fidelity to stick with us, year after year.
Okay, Resolution and Independence anybody, or Fidelity? No? If I give this talk again, I must remember not to includes jokes that depend on knowing the titles of some of Wordsworth’s better known lyric poems.
But back to Muscat: I recognised in John a visionary quality. More exactly, we all have visions, but John’s are useful. Let me illustrate this. I cannot recall the date, but it was the early years. I arrived at the office at the Innovation Centre, and John, in a state of excitement, said “Martin, you’ve got to see this”. He then took me over to the PC and showed me, in a somewhat earlier form, all those features of browser technology with which we are now so familiar: the url at the top, pages of information slowly pluthering down the screen, images, text, links to other pieces of text. There was even a webcam image that renewed itself every minute or so. One of the businesses there had rigged up a camera which took a photo of the scene outside their window and retransmitted it at regular intervals. I said, “what is this John?” and he said, “It’s a new system for linking together all the world’s computers and transmitting information between them in a simple and standard way. It’s called the internet.” (Or he might have said “the world wide web.”) I knew John’s capacity for over-excitement, and felt it was my duty to bring him down to earth. “John,” I said, “this will never catch on.” John does not create the future, but has an uncanny knack of seeing where it’s going, and finding an accelerated path towards it.
Eventually Muscat was taken over by the Dialog Corporation, but it all went a bit wrong and we both resigned. The problems were partly external: a recession loomed and the bursting of the dot-com bubble in 2000 brought down many of the good internet companies, along with the bad; and partly internal: the company structure was a holding company, Brightstation, that would spin off successful subsidiaries. But the subsidiaries fought over common resources.
(I truncate the long story. It is just business politics, and not interesting.)
Happily, when John and I argue these days, it is only over trivia, one being which of us resigned first, since we hadn’t planned to do it together. Here I have an exact timestamp which is against my resignation email,
- 12:24 and 32 secs pm, on Thursday, 22 February 2001
Then John and I went different ways but kept in touch. John started some new business ventures in Cambridge, I vanished into software writing. First came snowball, then Grapeshot. This time I had no doubt about choosing the name. Muscat was a type of grape, Muscat had ended up as part of Smartlogik, a subsidiary of Dialog, and Grapeshot was an aggressive way of saying that this was going to shoot Muscat down. Although of course I was really very fond of the Muscat software. So that is the real meaning of the “Grapeshot” name, although I was aware of it describing the scatter-gun approach to keyword matching. Grapeshot was written over the period 2002−3, during which time John‘s and my interests converged, and we again went into business together. Eventually Grapeshot Ltd was created as a holding company for the software, and another thing John and I like to argue about is whether I founded Grapeshot Ltd, or whether we founded it together. Well, I am in possesion of this document,
the Certificate of Incorporation, which I set up myself and which I keep jealously locked up in a filing cabinet in Norwich. It is to the Company what Magna Carta is to British Law, or the Declaration of Independence is the the USA. I take it out periodically and gloat over it like Gollum gloating over his ring.
But it’s time to let go. So without more ado I hand it over for safe keeping from now on to the person most deserving to take possesion of it. [Hands it to John.] (Don’t lose it: we might need it if ever we’re acquired.)
Finally, some other people in the Company who have been important in different ways,
Gael, who refused to let me chicken out of giving this talk,
Andy Morley, the last executive I ever interviewed or ever will interview. That was on Wednesday 1st July 2009. I didn’t listen to anything he was saying. I just thought, this man is a salesman, could he sell me dead horse? On balance, I thought if anybody could, Andy could, so he got my vote.
With Malcolm Cox, the other way round. It was Wednesday, 24th April 2013, and very chilly. He took me to a hard bench in Berkeley Square, and as it got colder and colder, plied me with questions. “good heavens,” I thought with alarm, “this is what an interview is like”. I barely knew, as I have not had a formal job interview since 1971. But if I failed this interview, I never found out.
Sioned Arrowsmith. You see them here, in wizard guise, locked in mortal combat with a balrog. In their left hand, Gandalf’s staff, in their right a phial of bitumen from the mines of Mordor. I think that’s right: it is a terrible photo. Sioned is a person of enormous importance, being the only one (with the possible exception of me) fully to understand the core Grapeshot software. I urge you all to keep up the pretence, at least, of wider knowledge. That applies equally to the people outside Cambridge. So if Andy Morley is asked what he knows about it, he should not say, “Oh, I leave all that to the technical guys at Cambridge”, but rather, “Oh yes! Martin has personally taken me through all 33,000 lines of the source code. Of course I can’t say too much about it, a lot of it is still pretty much under wraps, but I can tell you it’s pretty solid stuff”. It may seem that I am encouraging lying, but I think that is too absolute a way of thinking of it. I would prefer to say, “seasoning our customary discretion with a certain admixture of disingenuousness”. Let us not be too fastidious here. We are in the advertising sector after all.
Finally back to John. I’ve talked about our similarites and differences. One difference is class, John from the upper regions, me from the lower. Not that I myself am completely without pedigree: I can trace my genealogy almost as far back as my grandparents. But with John it’s a very different matter. Here is a typical John relation.
She is Sybil Sassoon, heir to an immense Rothschild fortune, Countess of Rocksavage, and later Marchioness of Cholmondeley. Notice the Snyder eyebrows! She lived here,
Yes, it’s her place, not John’s as I said before. Houghton Hall in Norfolk, built by Robert Walpole. According to the little guide book of Houghton which I have, she was “one of the most remarkable women of her generation”, and who am I to question such a judgment? She was certainly one of the wealthiest, and we may regret that the Snyder family tree must have been in some way skewed so that John failed to inherit a small fraction of it. For then we would have needed no external investment for Grapeshot, and board meetings would have been much more fun.
The portrait is by John Singer Sargent. Many of John’s relations were painted by Sargent, and you see John here, Sargent catalogue in hand, going in search of more images at the National Portrait Gallery.
And this is John’s best known close relation, though I don’t now recall the name. No Snyder eyebrows, and what looks like the Hapsburg jaw is, I have heard, the effect of ill-fitting false teeth. I think he did well for himself in America.
Any questions?