Skip to content

Notes

The second deployment, without the founders in the room

The second deployment is the first one done without the founders. It rests on what they wrote down and who they trained.

By Alessandro Singh and Rohith Varma, Updated , 8-minute read

In short

  • Count a deployment as the second only when the company's own team or a partner does it without the founders, in paid use after the pilot.
  • Know-how stays with the people who did the first site, so write down the integration kit, the runbook, the acceptance test, the training, the failure log and the cost and price per site while you work.
  • Decide who leads the second before the first one ends: a deployment lead from the first site, a trained partner, or founders on call and off site.
  • Record the time and cost of every deployment: each should take less of both than the last, and an investor should ask to see that record.

What counts as the second

Our thesis is that a technology company is proven at the second deployment. By a deployment we mean technology in paid use in a customer's own operation, after the pilot has ended. The second is the first one the company does without its founders. Our note on moving from pilot to paid deployment covers the first; this one covers the second.

One of our five tests, a way to deploy again, asks for exactly this: the company's team or its partners do the next site, business unit or country without the founders in the room, in less time and at lower cost than the first.

There are two kinds of second, and they test different things. A second site with the same customer shows you can deliver again. A second customer who pays shows you can sell again, and it is the one an investor will ask for.

The difference is more than a label. In a study of 36 pizza stores owned by 10 franchisees, know-how gained by doing moved between stores with the same owner, but not between stores with different owners (Darr, Argote and Epple, Management Science, 1995). Inside one customer, its own people carry the lessons from site to site. At a new customer, only your team can carry them.Source 1

Why know-how stays where it was made

The first deployment is often run by the people who built the technology, and what they learn there does not travel by itself. The information needed to solve a problem is often costly to move, so the solving tends to happen where that information is held (von Hippel, Management Science, 1994). He called such information sticky, and one remedy is to invest in making it less so.Source 2

The main barriers are not about effort. In 122 transfers of best practice inside eight companies, the major obstacles were the receiving team's lack of capacity to absorb the knowledge, causal ambiguity (uncertainty about exactly why the practice works) and a difficult relationship between the two teams, not a lack of motivation (Szulanski, Strategic Management Journal, 1996).Source 3

Results can also shrink when they travel, because the people and the situation change (Al-Ubaydli, List and Suskind, International Economic Review, 2020). In a study of six health technology programmes over up to three years, those that were complex in several ways at once rarely, if ever, became routine (Greenhalgh and others, Journal of Medical Internet Research, 2017).Source 4Source 5

Causal ambiguity is the founders' problem in particular. A founder fixes many small things quickly at the first site and writes few of them down, so nobody else knows why it works.

What to write down at the first

Treat the first deployment as the draft of the second. Write these down while you do the work, not after it:

  • The integration kit: every system, interface and data feed the technology touched, and what had to change at each.
  • The install runbook: each step, who did it, how long it took and what went wrong.
  • The acceptance test: what the customer measured before it said yes, and the result.
  • The training: what the customer's people had to learn, and the material that taught them.
  • The failure log: every fault, how it showed itself and how it was fixed.
  • The cost and price per site: what the site cost you to install and run, against what the customer paid.

Leslie and Holloway, writing on how startups learn to sell, describe the same work on the engineering side of a software business: finding out which features are unfinished, how easily the product installs and how well it works with new systems (Stanford Graduate School of Business, 2006).Source 6

The cost per site matters because another of the tests, a buyer who pays, asks for a price that covers the cost of delivering it. A runbook that shows the cost falling is the evidence.

Who runs the second

Decide who leads the second deployment before the first one ends. Three arrangements keep the founders out of the room:

  • A deployment lead who worked at the first site and owns the runbook.
  • A partner, such as an integrator, trained at the first site and paid to run the second from your runbook. A second site the integrator builds from scratch proves the integrator, not you.
  • Founders on call and off site. They answer questions; the team makes the decisions.

Grow the team after the learning, not before it. Leslie and Holloway found the same in sales: from 25 years of experience with startups and new products, they concluded that hiring a full sales force too fast burns cash and misses revenue targets, because the whole company first has to learn how customers buy and use the product (Harvard Business Review, 2006).Source 7

Investors look for the same sign at the customer. In 2024 Innovation Endeavors, which invests in robotics companies, set out how you typically know a company has real fit with its market: the product is used in production, and customers use it repeatedly without the company's involvement or expand it beyond the first scope.Source 8

Record the time and cost of each deployment

Two is where the proof starts, and it does not end there. Each deployment after the first should take less time and money than the last. Keep three numbers for each one: the days from signature to live use, the hours your people spent on site, and the cost against the price.

Learning is common, and it is not automatic. Organisations usually grow more productive as they gain experience, but some show little or no learning. Reasons for the difference include forgetting, staff turnover and knowledge carried over from other products and organisations (Argote and Epple, Science, 1990). In the pizza stores, unit costs fell with experience, and the knowledge gained wore off quickly (Darr, Argote and Epple, 1995). Knowledge you do not write down wears off, and it leaves with the people who hold it.Source 9Source 1

Some industries plan on the fall. In 2023 the US Department of Energy said repeat deployments of advanced nuclear reactors are expected to help cut their overnight capital costs by 40%, and called for committed orders for five to ten deployments of one design.Source 10

When this does not apply

Founders stay in the room for longer in some fields. In complex technical sales the founders often sell and deliver the first several deployments, and that is how the company learns. In those fields, the second deployment is the first one the company does without them, and what matters is what was written down at each one before it.

A second site can be a favour. A customer that already trusts you may agree to one to please a champion, and customers also buy for show: in 2023 the robotics engineer Benjie Holson wrote that one pilot can be all a customer needs to impress its shareholders, executives or the press. So ask who paid for the second, and from which budget. A second customer who pays is the stronger proof.Source 11

Some sites differ too much to share a runbook. Then group them by type, and count the second site of the same type.

Some technology needs no deployment at all. Software that people adopt one by one has no site to install, and this test is not for it.

Questions to ask

For founders planning the second deployment, and for investors checking one:

  • Who led the second deployment, and was a founder on site?
  • Is it a second site or a second customer, and who paid for it?
  • How long did the first and the second take, and what did each cost?
  • What was written down at the first, and did the team use it at the second?
  • Has a customer used the product again without the company's help?
  • What will the third need that the second did not?

If the answers live only in the founders' heads, the company is not ready for its second deployment.

Where to start

If you have one deployment, name this month the person who will lead the second. Start the runbook at the site you have, and record the time and cost of every visit from today. If you invest, ask for that record before you ask for the pipeline.

How we help

Client resultFounder advisory, 2026

30x pipeline: $1M to $30M in pipeline and $1.6M in revenue during 2026 for a B2B company in AI-agent security. As their go-to-market team, we helped hire three salespeople, set up the sales processes and systems, coached the team, provided market access and closed the sales together with their founders. We also helped them raise their Series A.

By November 2026 their own salespeople were closing without the founders.

For a way to deploy again, the engineering is yours. We bring the commercial half: the integrators and channel partners, the contract and the price for the next site.

Our work is done when your team runs the next deployment without us.

More notes

  • AI you can rely on

    Relying on a vendor's AI: the evidence to ask for

    In a 2026 survey, vendors supplied over half of UK financial firms' most material AI. Firms still answer for it: ask for evidence that repeats.

    For enterprises, investors and founders

    Updated

    , 8-minute read

  • Deep technology

    Selling to a public first buyer: terms that lead to the second

    The first buyer of deep technology is often public. Write that contract so it leads to a buyer who pays from an operating budget.

    For founders and investors

    Updated

    , 8-minute read

All notes

Newsletter

One email a month

Our newest notes, and what we are seeing in AI, machines at work and deep technology.

We use your email only to send the monthly notes. See our privacy notice.

Sources

In the order we cite them, with the day we read each one.

  1. 1. The Acquisition, Transfer, and Depreciation of Knowledge in Service Organizations: Productivity in Franchises (opens in a new tab)

    Eric D. Darr, Linda Argote and Dennis Epple, Management Science 41(11), , Accessed , Back to the text: citation 1 of source 1, a citation 2 of source 1, b

  2. 3. Exploring internal stickiness: Impediments to the transfer of best practice within the firm (opens in a new tab)

    Gabriel Szulanski, Strategic Management Journal 17(S2), , Accessed , Back to the text, where source 3 is cited

  3. 4. 2017 Klein Lecture: The Science of Using Science: Toward an Understanding of the Threats to Scalability (opens in a new tab)

    Omar Al-Ubaydli, John A. List and Dana Suskind, International Economic Review 61(4), , Accessed , Back to the text, where source 4 is cited

  4. 6. Startups Need a Special Learning Curve for Sales (opens in a new tab)

    Stanford Graduate School of Business (the school's summary of Leslie and Holloway's article), , Accessed , Back to the text, where source 6 is cited

  5. 7. The Sales Learning Curve (opens in a new tab)

    Mark Leslie and Charles A. Holloway, Harvard Business Review (a practitioners' article, not peer-reviewed), , Accessed , Back to the text, where source 7 is cited

  6. 8. Lessons learned the hard way: robotics (opens in a new tab)

    Sam Smith-Eppsteiner, Innovation Endeavors (an investor's essay), , Accessed , Back to the text, where source 8 is cited

  7. 9. Learning Curves in Manufacturing (opens in a new tab)

    Linda Argote and Dennis Epple, Science 247(4945), , Accessed , Back to the text, where source 9 is cited

  8. 10. Commercializing Advanced Nuclear Reactors Explained in Five Charts (opens in a new tab)

    US Department of Energy, Office of Nuclear Energy, , Accessed , Back to the text, where source 10 is cited

  9. 11. So you want to do robots, part 3: how to escape pilot purgatory (opens in a new tab)

    General Robots (Benjie Holson, a robotics engineer's essay), , Accessed , Back to the text, where source 11 is cited

Tell us what you are working on.

A founder replies within one working day.