What They Don

1
What They Dont Teach You in Graduate School

• Scott Dietzen, Ph.D.CTO, BEA Systems CMU CS
  92/ CMU Applied Math/CS 84scott_at_dietzen.com

2
Dedication And Acknowledgements

• This talk dedicated to the memory of Bruce
  Nelson, CMU CS 82
• http//www.cs.cmu.edu/about/development/aboutnelso
  n.html
• Why I volunteered (i.e., why I felt sufficiently
  presumptuous)
• Bruce help plant the seeds that inspired my
  entrepreneurship as well as my time on the dark
  side (working in marketing sales)
• Ive had some modest success as an entrepreneur
• Transarc ? IBM (Principal Technologist)
• WebLogic ? BEA (V.P. Marketing Ha!, CTO)

• Acknowledgements
• I plagiarized heavily from Bruces talk
• Borrowed some content from top-tier VCs
• Special thanks to Redpoint, Accel
• Borrowed some content from fellow entrepreneurs
• Special thanks to the WebLogic team
• And Paul Maritz (Lessons Learned at Microsoft)

3
The Missing Syllabus

• Technology adoption cycle
• The hurdles to commercial success
• Why non-technical issues matter
• Why technologists feel underpaid
• High-tech company organization
• Start-ups
• Venture capital
• Managing for success

4
Classic Technology Adoption Curve
Adoption
This is the best case scenario---all technology
aspires to be legacy! At any point on the curve,
new technologies may wash out (few make past
each phase)
Market Growth/Consolidation
Time
5
High Tech Efforts

• For success
• Laser focus on specific customersnot
  theoreticaland vivid understanding of pain and
  needs
• Clear understanding of needs hierarchy (what are
  they buying? Where is the pain/need in their
  priority stack?)
• Most typical downfall
• Cool technology, but insufficient business value
• Too high-level a value propositiondoesnt
  connect with new spending realities and specific
  pain points
• Unrealistic goals Time/money insufficient to
  reach demonstrable value

6
Why High-Tech Efforts Fail (Cont.)

• Miss technology sea changes
• Proprietary networks (SNA, IPX, AppleTalk ? IP),
  SQL (Unify)
• Java/J2EE (NetDynamics, Kiva, ATG, BroadVision,
  )
• Too much complexity for mainstream adoption?
• DCE, CORBA, SAA, ATM, Infiniband
• Web platform (Web services, J2EE, .NET) still at
  risk
• Weak execution
• Team comes apart or doesnt come together
• Focus wanes
• Not sufficiently responsive to change
• Ferocious competition
• 40 web application server vendors ? 3 viable
  today

7
Why Non-Technical Issues Matter

• In the real world, technology is the means to
  the end not the end
• To change IT in the real world, you need to
  make the business case to the technology
  end-user
• Define compelling value proposition
• Articulate the return on investment (ROI)
• There is a very high hurdle for doing this
• Think about receiving phone solicitations at your
  home
• This is why marketing and especially sales are
  so highly-valued

8
Why Do Technologists Feel Underpaid
The intellectual level needed for system design
is in general grossly underestimated. I am
convinced more than ever that this type of
work is very difficult and that every effort to
do it with other than the best people is
doomed to either failure or moderate success
at enormous expense. Edsger Dijkstra

• Because they dont understand the game
• Technologists often feel sales and especially
  marketing are trivial
• But in most companies, RD is lt5 even in
  high-tech RD it is generally lt20
• If this were not an optimal model, then the free
  market would correct it

9
Technology Company Organization
Corporation
Finance and Administration
Research (Huge Companies)
Engineering
Marketing
Sales
Services Consulting and Support
Manufacturing
Customers (Solvent Companies)
Advanced Development (Most Companies)
10
How Departments Contribute Value

• Marketing
• Defines, targets and articulates value
  (specifications and collateral)
• Engineering
• Refines and implements value (designs and tunes
  products)
• Manufacturing
• Physically creates value (products)
• Sales
• Creates desire for and demonstrates value
  (explains products)
• Service
• Preserves, transfers and enhances existing value
  (care and customization)
• Finance and Administration
• Administers the value-adding machine (company and
  customers)

11
Marketing

• Marketing represents the company to the target
  customer
• Strategic marketing
• Directs product strategy and often corporate
  strategy (in conjunction with company thought
  leaders)
• Owns the company business plan
• Product/technical marketing
• Pricing, channel strategy, competitive strategy
• Tactical marketing
• Public relations
• Demand creation (finding/qualifying new
  customers)
• Public face (literature, top-level web-site,
  events, advertising, etc.)
• Strategic partnerships
• Company and product naming/corporate identity
  (Ouch!)

12
Marketing/Engineering Interaction

• With respect to Engineering, Marketing
• Determines product lines and families
• Specifies individual products externals at a
  high-level
• Engineering owns the internals and the details
• Is usually responsible for planning product
  configurations, packaging, and documentation
• Sets prices, options, and schedules product
  introductions

13
Sales

• Sales represents the company to the customer (and
  vice versa)
• Actively approaches and cultivates new customers
• Demonstrates a products added value and helps
  customer internal champions overcome objections
• Is customers liaison to the company for
  pre-sales support, post-sales product
  improvements, and win/win business dealings
• Forecasts business bottom-up
• Revenue forecasts are prime corporate feedback
  loop
• Note Sales channels are typically very expensive
  (This is a fundamental start-up conundrum)

14
Marketing/Sales Interaction

• With respect to Sales, Marketing
• Forecasts business top-down (Sales goes bottom
  up)
• Selects distribution channels (in conjunction
  with Sales)
• Value chain can be complex (Value-added ISVs,
  SIs, OEMs)
• Some product requirements come out of channel
  model
• Provides sales tools to sell (transfer genuine
  confidence) company and product value proposition
• Literature, advertising, demos, benchmarks,
  competitive analysis, product training,
  roadshows, seminars, and so on
• Conducts sales training (nontrivial)

15
Life of a CTO

• Extremely inter-disciplinary role
• Spend time with
• Customers and partners via sales (50 in my case)
• Industry orchestration (Standards)
• Engineering (defining release themes and
  refereeing disputes)
• Corporate strategy/governance
• Evaluating start-ups/ISVs as partners/acquisitions
  
• Investors and financial analysts
• Industry analysts and press
• In practice
• Meetings, presentations, phone calls, and email
• Stretched a mile wide and ½ an inch deep
• Far more grunt work than deep thinking (worse
  than academia)

16
Comedy Break 1 Top Ten SignsYou May Be
Traveling Too Much
(All true stories)

• 6. You have to go to the hotel front desk to have
  them look up your room number (since its not on
  the plastic key)
• 7. Flight attendants greet you by name
• 8. Youve eaten nothing but airplane food for
  more than 24 consecutive hours
• 9. You read the hotel stationary to find out what
  city and country youre in
• 10. You have a nightmare about being on an
  airplane, and then you wake up on an airplane

17
Comedy Break 1 Top Ten SignsYou May Be
Traveling Too Much
(All true stories)

• 1. Your neighbor mows your lawn for you
• 2. A United pilot tells you youve been flying ½
  as many miles per year as he does
• 3. You work a full Wednesday in Japan, fly to the
  states at the end of the day, and arrive in time
  for work 10am Wed. morning!
• 4. A constant state of sleep deprivation ensures
  that you instantly adjust to any time zone
• 5. Not only cant you remember where you parked,
  but you dont even know the make, model, or color
  of your rental car (worse in a blizzard)

18
Why Do Technologists Feel Underpaid (Redux)

• Because they dont understand the game
• Responsibility for product and company success is
  paramount, and rewarded accordingly
• Accurately defining the product for the target
  market is generally much harder than building it
• Domain expertise is often more valuable than
  programming skills
• Working with others (leading and managing) is
  harder than programming
• Leaders are most highly valued, then managers,
  then individual contributors
• Asking for money is hard orchestrating a
  technology purchase can require months of grunt
  work and negotiating enterprise licenses can be
  excruciating
• As a result, the biggest bucks go to the
  management team and to sales people on commissions

19
Start-Ups

• Successful high-tech start-ups (generally)
• Identify existing target market
• Have channel of distribution to that market
• Have unfair technology advantage
• Have time to market advantage
• Have tremendous growth potential

• Start-up employees
• Are freed from maintenance/legacy
• But must work very, very hard
• Are usually substantial stock holders
• Have to wear multiple hats
• Must be able to work well on a team

20
Evaluating A Start-Up

• Evaluate from an investors view point
• Review the business plan, not just the technology
• Corollary Avoid the technically interesting, but
  commercially unviable
• Ask hard questions about risk and competition
• Under non-disclosure, get as much detail as you
  can about financing structure (this may be
  tough!)
• Meet with angle investors/VCs if you can
• Most of all Do you love the team, love the
  technology, and have fire in your belly!

21
Joining A Start-Up

• Equity considerations should be primary
• Expect to take a significant pay cut over what
  you could make elsewhere
• Translate equity offers into percentage of the
  company (Shares outstanding range from 8m to
  20m)
• Founders should expect single digit percentage
• Lead founder often in double digits
• CEO 5
• Management team 2
• Senior engineers lt1(even early on)

45 Investors
20 Founders/ wizards
10 Management team
15 1st year employees
10 Later employees
Fractional Ownership of an Typical 1st Round
Tech Venture
22
Premier Venture Capital

• Why?
• Defacto seal of approval for company/product
• Trade share of ownership for resources
• Let the pros finance the money-losing first
  years of company development (Your risk should be
  opportunity cost)
• Get into the network
• Early warning about competition
• Help with recruiting the management team
• Easier to get money down the road
• Smarter than industry money

• Why not?
• Dilution
• Interference/loss of control

23
VC Profile of an Early Winner
24
Venture Capital Home Run Illustration

• WebLogic Funding
• Employee option strike price starts at zero
  (founder) and goes up to .33/share
• Angel financing Regis McKenna, Frank Caufield,
  Ali Kutay
• First (and final) round in 97 12.5m in venture
  and strategic industry funding
• VCs TL Ventures Bay Partners (Both got board
  seats)
• Industry Intel and Cambridge Technology Partners
  (SI)
• Sale to BEA 10 months later
• Approximately 10X return on investment (home
  run)
• Grand slam if they held (BEAS goes from 25 to
  350, split adjusted)
• BEA Funding
• First (and final) round in 95 50m from Warburg
  Pincus!
• As most of the limited partners held, their ROI
  reached 100X

25
Current VC Environment

• Conventional wisdom
• Back to basics
• Dumb or bubble money has left the market
• Kind of like the early 90s
• Reality Nothing basic about pounding hangover
  from the excesses of the late 90s
• Major overhang of companies funded in the boom
• Still too much capital flooding the obvious
  categories
• History suggests that the downturn could be much
  more prolonged

26
2002 So Far Like Its 1998
Deal Flow and Equity into Venture-Backed
Companies (annual)
Amount Invested (B)
Number of Deals
Source VentureOne
27
80 Of Dollars Directed At 2nd And Later Rounds
Source VentureOne
28
Environment
Accumulated Capital Over-commitments (B)
29
Venture Capital In 03

• In 2000, 653 venture funds raised 107B
• In 2001, 331 venture funds raised 40.7B
• In 2002, 108 venture funds raised 6.9B
• Moreover, 26 firms gave 5B back to their limited
  partners
• So net new 1.9B go into venture funds in 02
  (95 drop!)
• Total over-committed capital for all private
  equity is 100B!

30
Technology Investment Cycle4 Phases
1967-1969 Semiconductors
1970-1971
1981-1983 PC Workstation
Bubble (2 yrs)
1984-1985
Free-Fall (2 yrs)
1996-1999 Internet
2000-2001
Firming (5 yrs)
1961-1966
1972-1974
Bottoming (3 yrs)
1975-1980
1986-1989
1990-1995
Today
31
BottomingCompany Hangover
From 1995-2000
14,463 978 1,529 1,180 10,776
Companies funded Went public Were acquired Went
out of business Remaining
Mortality effects, me-too effects still working
their way through the system
Source Venture Economics Venture Source
32
Internet Investment SettlesAt Early Levels
Equity Investment by Internet Dimension
Amount Invested (B)
Source VentureOne
33
Industry Allocations Fluctuate In 2002
Source VentureOne
34
Software Clearly Dominant In 3Q02
Percentage of Overall Venture Investment, by
Industry
Source VentureOne
35
What VCs Want

• VCs generally are not anticipating disruptive
  technology (e.g., internet, optical networking)
  revolutions near term
• Broad, hugely ambitious plays are out More
  realistic narrowly targeted (segmented) plays are
  in
• So dont get too enamored with technology
• The game today is applying technology to a
  business problem that customers will gratefully
  pay to solve!
• No more Just get the eyeballs today, and figure
  out how to monetize later
• Corollary Technology is only a small part of the
  business plan. Compelling marketing and selling
  strategy are essential

36
What VCs Want (Cont.)Funding to Milestones
Idea isFeasible
Model is defined and proven
Beta or FCS
Valuation
Risk (ß)
Capital
Seed
RD
Ship
Expansion
37
What VCs Want (Continued)

• Make the business case What is the customers
  return on investment (ROI) that will allow you to
  defy gravity
• Industry/domain expertise on management team is
  mandatory!
• Strong syndication
• Money was a commodity in the bubble. No more
• Other smart, well connected VCs in the deal
• Angels also key
• Shared risk or herd mentality?
• Entrepreneurs need multiple VCs on the hook!

38
Mind of a Successful Entrepreneur

• Today (Reality)
• Survivor
• Milestone-based spending, lag behind the curve
• Get (named) customers!
• Cash flow breakeven is key (min. team, share
  hotels, gear on EBay, )
• Raise enough to get to a significant milestone
  ROE is a function of tangible business progress
• Protect the downside

• Late 90s (Bubble)
• Go big or go home
• Build infrastructure, spend ahead of the curve
• Get buzz
• Time to market is the key
• Raise only the minimum required ROE is a
  function of time
• Optimize for the upside

39
Start up/VC Summary

• Despite the economy, major shifts continue to
  occur
• In many ways, great time to be starting a company
• Costs are down dramatically
• Talent is available
• Luxury of being out of market scrutiny
• But, caution and prudence still rule the day
• Industry exceedingly well funded
• Business fundamentals not a V recovery
• Experience matters

40
Comedy Break 2 Sabbatical Actually Helped My
Career!
41
Managing For Success

• So youve got money in the bank. Now what?
• Three golden rules
• Cash is your life supply, manage it ruthlessly
• Get the first version out the door quickly...
  don't wait for the perfect product 
• Bad decisions happen recognize them, fix them
  quickly, and move on 

42
Managing For Success I

• Take great care in team building
• For start-ups/new product launches, only small
  teams of world-class talent will do
• Ensure critical team roles are filled
• Architect Sweats the content (big picture and
  details)
• Manager Sweats the process
• Product manager Customer/scenario/solution
  champion
• The Cheer leader/Ass-kicker
  Demands/inspires excellence and commitment
• Some overlap possible
• Success depends upon excellent teaming among four
  roles

43
Managing For Success II

• Establish key scenarios/priorities to guide
  decision making around product development/release
  s
• Product releases should have a small (2-4) number
  of themes
• Doable in reasonable timeframe
• Validate themes with customers
• And then get ruthlessly and relentlessly behind
  them
• Use scenarios to allocate resources
• Use scenarios to organize engineering
• Push ownership down as low as possible (to the
  natural scenario owner), but no lower
• Focus, focus, focusHigh-tech projects can
  easily get distracted (the earlier in the
  life-cycle, the greater the danger)

44
Managing For Success III

• Acquire with extreme caution
• For people and IP? or
• For market/customers?
• Know who is going to manage it beforehand
• Carefully balance
• Assimilation
• Preservation
• BEA has done 30 acquisitions, but only a couple
  major waves of change (Built Entirely on
  Acquisitions)
• Tuxedo provided initial engineering culture
• WebLogic first major agent of change
• CrossGain/WestSide second major agent of change
  (still running its course)

45
Start-Up Lessons

• BEA/WebLogic won because some big bets paid off
• Web would push more business logic to server-side
• Java was a great language for server-side
  programming
• Java on the client would drive Java on the server
  (Mostly wrong!)
• Java VM architecture would prove to be
  dislocating agent for existing OLTP
  architectures (Tuxedo, CORBA)
• Need a API standard for the server Enter J2EE
• Bundled suite of platform services ? Web
  application server
• Free trial web download (over sales objections)
• The jury is still out on BEAs future big bets
• Can we continue to carve out a niche between
  Microsoft and IBM?
• Can we extend web/Java platform to encompass
  integration EAI, B2B, UI (portal), data (XML
  Query)

46
Whats Hot Web Application Platform

• Complexity of programming business applications
• Complexity/performance of XML processing
• Semi-structured content repository meets file
  system and database
• Complexity of application integration
• More semantic models of XML data
  mapping/transformation
• Choreography/orchestration missing
• Complexity of operations, administration, and
  management
• More adaptive (self-configuring, self-healing)
  systems to meeta quality of service
• Business activity monitoring (BAM)
• Security infrastructure (may be oversubscribed)
• Wireless (historically oversubscribed, still
  working off excess)
• Data center consolidation on disposable hardware
  (may be oversubscribed)

47
Recommendations forThe Would-Be Entrepreneurs

• Can you be a BSO (blind, stupid optimist)?
• Learn how to write and speak well
• Learn to be a consummate team player
• But also be a leader Step up and take
  responsibility
• Spend some time working on the technology
  transfer/business side (technical marketing,
  technical sales, consulting)
• Consider auditing classes in entrepreneurship at
  GSIA
• Learn to be comfortable making decisions from
  incomplete information
• Have some fun now (while you still can)

48
Bruces Final Oral Exam

• Purely technical issues are of tertiary
  importance. Predicting technology evolution and
  delivering high-added value to the business are
  first and second
• CS graduate school most typically blissfully
  ignores this
• Start-up companies have the best wealth
  potential, but require the highest effort,
  discipline, commitment, and risk
• The greatest technical influence comes out of
  connecting technology to the business (technical
  marketing, product management, CTO, etc.)
• Compensation is commensurate with responsibility
  and ability to assimilate incomplete information
  and act decisively upon it
• Many Ph.D.s with maturity become great technical
  managers. Far fewer become great marketeers or
  executives

49
Thank you!

• Value-added contributionsmost welcome
• (If this can be made sufficiently valuable,
  then Id like to do it annually)