Science and TechnologyWritten evidence submitted by Matthew Bullock

Starting a technology business without venture capital funding has been a familiar challenge in the Cambridge area since 1960s, but the Phenomenon there was and is still largely not funded by venture capital, although the periodic availability of venture capital is welcomed as additional funding for growth.

It would thus be wrong to regard the technology pathway as blocked at the outset because of a lack of venture capital funding. It is always the “first best” solution for companies if it can be raised, but the number of companies that get venture capital will always be a small minority, and too little time is spent examining the “second best” option adopted by the great majority, and whether there are steps that could be taken to improve that.

In my experience, venture capital more commonly comes in at the later stages of a company’s development, building on a market and product position that has been developed more gradually along the alternative route, and then kicks the company’s growth trajectory onto a higher plane.

What is the alternative route and how does it come about? How is it financed? And what could policy do to improve it?

I christened this route the “Soft Company Model” in 1984, to distinguish it from the “Hard Company Model” that I saw had come to be favoured by US venture capitalists after around 40 years of predominantly soft company growth in the US.

The soft company model grows out of the nature of most technology sales: these predominantly involve the delivery of an intermediate industrial component or system to a large lead client—corporate, institutional or governmental; there are very few technology companies that sell to consumers, at least in the early stages. The sale is usually, therefore, based on an R&D or technical equipment contract for delivery over a period with agreed milestones, often with stage payments. It is not a sale from a catalogue, with simple payment on delivery.

The product is frequently quite specific to that lead client’s needs and is based on the founding technologist’s understanding of how to apply a broad corpus of technical expertise to a client’s particular problem. This problem solving of specific issues often starts off as a consultancy service, or offering a testing or design service. The company may then build a one-off system that embodies their design skills. This can then transition to a more routinised delivery of their service, before the company reaches the point where it feels sufficiently confident from its own experience to launch a discreet product with defined features that will address most of the demand from its clients.

How quickly this happens depends in part on the evolution of the technology. Most initial science in a new area is quite inchoate; a lot of initial work goes into measurement and other instrumentation as the problems get more defined. In these circumstances there is a risk in rushing to “reify” issues and solutions too early; the smarter move is to place a series of lower risk bets around solving particular problems with the developing technology, than to bet that Product X is the solution to all current problems in the field.

This gradual product “Hardening Process” can take a long time to achieve: in TAP Biosystems it took us several years before we and the market were ready for standard products; in other technologies a clearly defined product need may emerge over two to three years. Hardening may also be paced by the internal development of the company: in TAP’s case a great many internal changes were necessary to alter our processes appropriately after we decided to manufacture standard products, following years of designing and building customised systems.

This progressive approach also sits more comfortably with the development of the scientific entrepreneur’s own personal skills. In my experience scientists can become better than average business people once they have applied themselves to it, but it takes time and experience to learn, even for quick learners like them, and the risks of building “softly” are much more manageable than plunging into a “hard start”.

The financial risks are quite different: assuming the research costs have been incubated elsewhere, the first costs will be the salaries and overheads of development work, plus such equipment as is needed for this; in a “hard company” start up there will then be expenditure on manufacturing equipment and staffing, distribution and sales staff, and marketing expenditure to launch the product; on top of this will be stock/work in progress and debtors, maybe partially offset by trade creditors. The rule IBM adopted for a new product launch was 1:3:10: 1 to stabilise the product; 3 to get it manufactured; and 10 for marketing. That would not be untypical for a novel piece of equipment; it would be substantially greater for a novel pharmaceutical product, because of the clinical testing required. And for a start up, all that has to be incurred with an untested management team, no certainty of customers or sales at the end, and losses increasing each month until the launch is successful.

For a “soft company”, the development costs are the same, but the manufacturing costs are lower and the marketing and working capital costs are much lower; the processes are more familiar to the management team; crucially, the costs are primarily incurred against a certain sale to a credit-worthy client; and, depending on the contract terms, the company can be making monthly profits and getting stage payments as it reaches the agreed milestones.

Moreover, soft companies are usually quite profitable from the outset and, once they have a better feel for the market, they can use their profits to invest in the internal development of a hard product idea, with which to kick up their subsequent growth rate; in my experience the resilience of the soft company model meant that several attempts at hard product launches could be tried without wrecking the business; and often it was at the successful launch of such an internally incubated hard product that venture capital could be brought in.

As a result, while the first company’s business plan has to be financed by equity, the second, soft company is usually financeable by a bank with a closely monitored, secured working capital facility from early on in its life, and with personal equity and/or a government loan guarantee scheme, even from start up. This was how Barclays was able to support the growth of Cambridge companies well before the availability of venture capital in the late 1980s—as indeed the Bank of Boston and Bank of America had done in Boston and California in their pre-venture capital days.

What policy changes would make this Soft Company Model easier to start and succeed?

An important point to make is that the development of technology products and equipment on contract—what the great, soft majority of smaller technology companies do—does not appear anywhere in the R&D statistics: the activity does not conform to the Frascati definition used by BIS/ONS to measure R&D activity; for the small supplier the activity is recorded as sales; and in their large customer’s accounts it may appear as capital expenditure or revenue expense. The Frascati definition requires the activity to be speculative, without a firm sale in prospect—ie to be crossing the Valley of Death.

Since no one measures soft company activity, there are no tax allowances or other incentives for it.

The first suggestion would be to develop a method of measuring the activity.

The second suggestion would be to encourage the growth of the market for technology procurement from small companies by larger companies, research institutions and government departments: development contracts in my experience are much more commonly available from US corporations as a way for them to develop access to new technologies than in the UK—so-called “open innovation”; in particular, British government procurement has long operated a closed innovation system for procuring technology in defence and has shown remarkable unwillingness until recently to consider using its procurement capability in an open way to seed new technology or policy areas—no doubt because the civil servants fear being arraigned before the Public Accounts Committee for wasteful expenditures. Contrast this with the very effective open procurement activities of such agencies as DARPA, ARPA and ONR in the US, which have been used as a conscious stimulus to advanced technological and economic growth since the 1940s. The Maddock Report made a similar suggestion in the UK in the 1970’s. More active funding of SBRI contracts in government departments, agencies and the Research Councils would be a first step in this direction.

A further suggestion is to revive the pre-production prototype financing scheme that was once offered by the DTI. This was available to large UK companies that placed prototype development contracts with small technology suppliers to cover the cost of experimental pieces of equipment. Adapting the TSB’s multi-partner collaboration R&D mechanism to fund bilateral partnerships between small technology suppliers and large companies would help the latter to move towards more open technology procurement models.

Lastly, a word in the ears of some of the banks to get back to doing more of this sort of lending would be useful.

April 2012

Prepared 12th March 2013