Semiconductor IP industry - ready for restructuring?

Part two: the challenges

Following on from part one, and in order to examine some of the current challenges facing the semiconductor IP (SIP) industry it is worth looking at the genesis and evolution of a typical tech-based/service-based start-up. The following scenario, which might be a typical source of the industry's current malaise, illustrates a few of the challenges.

A few friends (Michael, Terry, Dave - MTD) are frustrated at their current company and decide to bail out and start their own operation. Not having much commercial experience or market knowledge, they offer their current skill set to their friends who work for Company A. In other words, the origin of a SIP company is a 'body shop', which sells a design service capacity built up over the last few years. MTD fills in the gaps for a larger company, which at a particular point in time either lacks sufficient internal resources or the skill set to finish a given project in a cost effective and timely manner.

So MTD are lucky and finish the project within a few months and build sufficient knowledge about their clients internal operation, hurdles, bureaucracy, management style, and figure out that these problems are common amongst most large companies. They approach Company B and offer to provide the same level of design service as for Company A. MTD quickly figure out that 20-40% of the work required is a repeat of the work that was carried out for Company A. They can keep quiet and risk copying the design that was already done for Company A, but realise that this will soon land them in trouble, so they ask Company A for permission to re-use some of the design, in return for discounted fees for the next project. Company A agrees, as they cannot see much value in the old design, and cutting the cost of the next project will reflect well on the management team.

MTD find that this particular design problem is common to most of their customers. By using higher priced license fees, they can considerably reduce the additional charges for the body shop portion (which incidentally, is funded by the customer from a different budget) and undercut potential competitors.

So next time MTD offer their design services, they can shave off considerable time for their clients by using pre-designed blocks, but they will need to charge a fee for these pre-designed elements. So now we have the birth of the licensing business, and let's call the pre-designed blocks intellectual property (IP), so that others may not copy it without paying for it. The more customers MTD engage with, the more knowledge is collected and the more pre-designed blocks are inventoried, tested and documented. MTD is now in the position of offering a large number of pre-designed blocks, which are now sold as SIP.

The fee paid by the customer for support is now classified as maintenance. So far so good: your SIP saves the customer considerable time, resources and, since they can buy this particular knowledge in the open market, they no longer have the need to maintain this knowledge base within the company anymore. So the customer develops a deep engagement with their SIP supplier. The SIP supplier can smell the dependency, and use this knowledge to continue to squeeze higher license fees and now demand a bigger share of the proceeds. As it is too late for the customer to go back in time and reinstate the team of internal experts, they grudgingly agree to payment, based on the total number of products that will manufacture and sold. In return for supplying the SIP and related services, MTD will now receive royalty payments when the customer puts the system using MTDs SIP into production.

But like most new start-up sectors, MTD and their compatriots face some new challenges. We have summarised a few of these below:

1. A black box set of pre-designed elements called SIP is OK, but when you interface it within the customer's system, the box has to be modified, retested, and documented. The black box then becomes a shade of grey. But MTD still faces the challenge that their clients cannot easily integrate the pre-designed blocks into their overall system design and need a lot of support and manpower to finish the project. MTD will need to charge far higher fees next time round, as they had under priced the cost of this extra support.

2. Each customer is using a different design methodology, and a different set of design tools, from many different design tool vendors. MTD is now expected to have sufficient knowledge of not only the customer, but also their methodologies and tools.

3. Each design department of the customer has different budget authorities and limits, often with no authority to negotiate or sign a legal agreement without consulting internal or external legal entities. The accounts department is not keen on the paperwork, administration, and calculation methods used for royalty payments.

4. Customers have their own pre-approved brand of Non-Disclosure Agreement (NDA), which is different to the SIP vendor's NDA. How does the SIP customer protect both their existing knowledge, and the SIP vendor's property, from leaking within a company? This needs to be worked out before the company decides to license the SIP.

5. Each new contract takes increasingly longer to negotiate as lawyers, accountants, and more company executives get involved in the deal.

6. Once the customer is engaged, new, previously unidentified factors come to light, which require the SIP vendor to spend double the resources that were allocated for this project. Not wanting to upset an important client, the SIP vendor continues with the support, hoping that the design will soon go into production.

7. The SIP vendor has very little influence on whether a design will go into production, or what quantity will be manufactured. It is almost impossible to forecast the likely quantity.

8. As more SIP vendors enter the market, offering the same commodity SIP blocks, MTD decides to specialise. The question of which area to specialise in is not easy to answer, as most of their resources are located in one country, but the customers are located in another. The target market has a different speciality than the one the SIP vendor has been focusing on.

9. Whether or not to offer a new specialty that will require a complete re-skilling of the design team, or even replacing some of the existing team, is a daunting challenge, which wasn't envisaged.

10. MTD has developed a third party network, some of which has helped MTD get established with complementary products, but other members have been badge collectors, who are dependent on MTD to provide them with new customers and require far more support than a typical customer before they expend any effort. Each third party is just like a customer, who will pay in the long term. The whole way of looking at third party networks has to be re-evaluated.

11. Meanwhile, some competitors have beaten you to the post, by offering a complete solution, which is still in the planning stages at MTD. There is a likelihood that you may lose your most important customer to yet another new start up. So you decide to take the risk (The EDA risk: you tell them what they want to hear, ship them what you have, and any problems that the customer discovers, will be fixed later, usually in the next release). Although this model, will save you a few customers in the short term, it may come to haunt you sooner than you think, as this habit (also known as generating a costly legacy) becomes part of the MTD culture, and soon every customer is getting the same treatment. Remember; customers talk to each other and bad news spreads fast.

12. The customer finds out that you never had what you promised, and threatens to cancel the contract and go public on this. You manage to overcome this by throwing more resources on it, and guess what, your future products (and the future of your business) are on hold, as you try and appease your limited number of customers. You may have seen this movie before. It was called, the "Body Shop", but this time you are not getting paid for the time, just the one off license fee.

13. If you had remained a body shop, you should have been able to garner approximately $100K to $150K per head per annum, but at the current rate you are not even making $50K a head from your current license fees. One big advantage of the body shop model was that any spare warm bodies that weren't generating revenue soon became conspicuous, but now with the SIP model, many of these bodies hide within larger teams, finding plenty of "work" to meet their curiosity levels.

In part two, we have looked at some of the challenges facing the SIP industry. In a downturn, companies are burdened with surplus resources, which they can shed or maintain for the upturn; when the upturn begins, the semiconductor industry will be driven by a whole new set of parameters. Part three of this series will examine some of these new parameters and suggest some of the options. We will also examine the different solutions, options and likely scenarios that are available to SIP vendors.

Comments on this story? Send an email to Picasso: picasso@thechilli.com

Picasso

18DEC2002