Tech Foundations for Startup Founders
As you build your startup, get the foundations right for your technology, processes, and people — or wing it and hope you haven’t built on sand, advises Programmer, Architect & Founder, Kevin Fleming.
This is the first article in a three-part series for startup founders without technical experience. The next two articles address the top 10 technology-related pain points experienced by startups at every stage.
If your core product or service is delivered via the web, a mobile device, or desktop app, you are actually using the definitely-not-sexy business/technology pattern for information manufacturing, distribution, and sales.
Over the coming years, you’re going to build a machine to do all that manufacturing, distribution, and sales, using complicated technology tamed by people I prefer to call tech wranglers, (rather than techies or IT people). Together, you’ll lay the foundations for the first four stages of the machine. With your unparalleled leadership, a great product, and some amazing people, your vision will materialize before your very eyes, delighting your customers, your team, and your investors — if you’re one of the ten percent of startups that survive.
So easy to get wrong
Before we hit the street, there are some other numbers I’d like you to consider that demonstrate how easy and costly it can be to get the foundations wrong, whether you’re a startup or a massive multinational:
Data-driven Business Failure Rate
- 85% of big data projects fail (Gartner, 2017) 87% of data science projects never make it to production (VentureBeat, 2019), reports Data science Process Alliance
Digital Transformation Failure Rate/Cost
- 70–90% of digital transformation projects fail to meet their objectives, according to various studies.
- $2.3trillion wasted globally in failed digital transformation programs in 2o23, says Taylor & Francis Newsroom.
Satisfaction with IT departments
Cloud Cost Surprises
As a founder myself, I want you to do more than survive, I want you to thrive! With three decades of experience as a programmer and architect (solutions, chief, enterprise), including four startups, a few Fortune 500s, and an R&D company solving a complex problem for the US Navy, I have prepared a business-technology architecture (in plain English) to help you execute your vision with confidence. As an added bonus, the information below will serve as an excellent set of topics when you’re interviewing a technical co-founder or key hire.
Foundations… for what?
A foundation, even for a technology company, is a simple structure. Its key requirements are coverage of an area large enough for the eventual people, processes, and workloads; it must be stable and strong enough to support future activities; it must be level and smooth enough to get the work done quickly, and at a reasonable cost. Most important, is that the foundation suits the vision, mission, strategy, and the plan.
The foundation for each stage is built with one or more of the following items in your toolkit:
- Data, information, and knowledge.
- Product-market fit (PMF)
- Customer journey maps
- The operating model for your machine
- Knowledge management system
- Management communication bus (MCB)
- The definition, design, implementation (buy vs build), and operation of the machine
- Short-term technology wrangler consultant/contractors (I dislike terms like IT people or techies)
- Contract-to-hire technology wranglers
- A technical co-founder and or key hires
- Technology strategy
- Technical debt
- Awesome code
- Other people’s software (OPS)
- FinOps
This article is concerned with the first four business-technology stages:
You could tack on another one or more stages, perhaps plateau, major pivot, diversification, acquisition, unicorn, IPO, exit, and so on, but four stages will serve you well enough for the next few years.
For each of the four stages, there are objectives and key results (OKRs) for the foundation being laid during that stage.
Put the foundations and OKRs together in a sweet infographic, and, slightly abridged, it looks like this:
Overarching Concerns
In addition to vision, mission, and strategy, there are a few more corporate keystones that you must put in place. Why have I included them? In my experience, they have been noticeably absent in almost in every workplace, to the detriment of those companies. When they have been present, they illuminate every situation, quicken decision-making, and keep the work on course and speed.
Core values and priorities
Corporate values are the fundamental beliefs and principles that guide a company’s culture, decision-making, and actions. They are the heart and soul of an organization, shaping how employees interact with customers, colleagues, and the community.
Priorities are obvious and may seem mundane, but in relation to values, they are critical. If a core value is knowing your customer, but you don’t prioritize collecting, sharing, or using that information, then it’s not much of a value; team members see that and act accordingly.
Technical or business requirements — the what — are never enough. Values and priorities are the why and when and how much/how hard/how fast.
Be data-driven from day one
Machine learning-based point solutions will improve by leaps and bounds over the next five years. You must take advantage of some of these solutions. You can only do that if a) you have high-quality data with solid metadata/provenance; and b) it’s democratized and understood to mean the same thing to everyone, from CEO to PO to Data Analyst. Look to your knowledge management system to do the heavy lifting.
Knowledge Management System
I’ll write more about this below, but I want to emphasize now that the KMS is one of your essential pieces of kit that has to be a core value and top five priority from day one.
Speaking of being data-driven, make sure you track who is contributing to and reading the KMS. This data is a perfect use case for a knowledge graph that can show you your influencers, and committed users, as well team members who haven’t taken this core value to heart.
Management Communication Bus
The MCB is a term I believe I have coined that is meant to criticize the usual rails for communication within organizations. Until recently, there was a somewhat popular technology called an enterprise service bus (ESB), which was a means of integrating multiple, diverse services. Basically, you’d dump some data onto the bus, and at one or more places, data would be shunted off to some code to process it in some way and then delivered… somewhere. Many tech wranglers loved the ESB; many, including me, hated it. After a couple of years in operation nobody had any idea who wrote any of the processors, or their intended purpose. In many cases, nobody understood the data scooting around the bus because it was changed by any number of processors, and/or the data provenance was lost in the mists of time.
To me, this operates very much like internal corporate communications, thus the name management communication bus — a flawed means of integrating multiple, diverse minds within an organization.
Example: The CEO says to a middle manager, who emails the team manager who adds an approximation of what the CEO said to the team chat without including the context that was never mentioned in the first email.
Kan you think of a better communication Management Solution?
Business Resilience
Business resilience for a technology startup is about building the capacity to detect, withstand, recover quickly, and learn from disruptions of any sort. It’s crucial to integrate resilience into the core of your business and technology strategies. Resilience cannot be achieved overnight, or purchased from Amazon, so, over time, build a strong culture with resilience as one of your core values and top five priorities.
And now, without further ado, your stages!
Stage: Product-Market Fit
Foundation Objectives: PMF is a primary, enduring process; create the operating model; seek, collect, democratize, and operationalize customer opinion and pain point data.
Key Results: % would buy when your product exists.
Tools: Customer journey maps, feature descriptions, pitch decks; road map; data collection: customer feedback, surveys, focus groups, friends, family, investors; MCB; first levers of the machine; create the KMS.
Core Value: Knowing and satisfying your customer (guessing not allowed).
The core feature of the PMF foundation is finding out what the customer wants, documenting and democratizing that knowledge, and then satisfying the customer’s wants in your product deck and road map.
You will formalize and preserve this foundation through the creation of the company’s operating model, including the value chain of product activities, and the beginnings of a knowledge management system (KMS). The KMS will link to customer journey maps, feature descriptions, and the results of data collection. You must keep this up to date and stocked with current metadata (created when by whom, how collected, etc.), which is essential to maintaining data quality.
Investors will want to see the key results. The knowledge management practice that you start now is the evidence that you take PMF as seriously as you claimed for the money they will invest in this pre-seed stage.
N.B. Google Workspace Docs is not adequate.
Stage: Prototype & v1.0
Foundation Objectives: A minimally functional version of your product or service; collection of user experience (UX) information and product NFRs.
Key Results: NPS, Churn, etc. Product UX telemetry, NFRs‡.
Tools: Minimally viable technology; technical debt; customer feedback; NFR & UX telemetry; operational data; social media and user community forum; topic & sentiment analysis; levers of the machine.
Core Value: Everyone knowing what the customer experiences, always.
‡ Non-functional requirements, e.g., page load time, key transaction response times, mean-time-to-detect (MTTD), and mean-time-to-repair (MTTR).
This is your first attempt at creating the machine. You will throw this away when you start v2. You will formulate a technical debt strategy, that makes it easier to make decisions about where to take on debt, tracking it so it can be watched for sudden increases in carrying costs (bugs, poor UX, etc), and then ensuring the debt does not reappear in v2.
The customer gets a consistent experience, as promised.
Investors gain confidence in your product, your commitment to your customer, and your ability to execute.
The major risks to this value are that it gets lost in the day-to-day business of getting product out the door and that it will be deprioritized by short-term concerns.
You will formalize and preserve this foundation through the inclusion of collecting and democratizing customer experience data, from all tools, in your operating model, value-chain of product activities, and knowledge management system. For example, new product features or change definitions include user journey/user experience expectations. Automated code generation makes this much less onerous for developers. These expectations then become requirements for developers, and acceptance criteria used by QA and product owners (PO) during the user acceptance testing phase (UAT). Data are available to all product team members. Key team members watch and share these data daily.
Stage: Serious Traction
Foundation Objectives: Rapid delivery of features on your road map; increasing quality; retiring v1 and technical debt; high-quality v2 code; resilient machine.
Key Results: Reductions in delivery estimate error, defect rate, regression rate, MTTD/MTTR, abs % of v1 code; increasing unit test coverage, QA and UAT test pass rates
Tools: Automated code generation; project/sprint delivery data; code analytics: code quality, retired v1 lines of code (LoC); automated testing; observability/visibility; levers of the machine.
Core Value: Rapid delivery of high-quality products to your customers.
At this stage, you have your customer’s attention and interest. Now you must increase their desire and convince the next wave of customers to buy (see AIDA) with the features they wanted that you now have. As traction grows, your potential customer base, along with the number and nature of features they desire. To capture these customers, and to fend off competition, you must be able to rapidly deliver these features.
To do this, you upgrade the machine. You need to start to retire v1 and replace it with much more robust (technical debt-free) code, architecture, and processes. This stage will also set the groundwork for the next stage, where scalability in team, practices, and technology will be key to whether you thrive or merely survive.
The customer gets a reliable product, upgraded per the road map, that is now part of their personal or professional lives.
Investors get the data you have collected that aligns with or exceeds your estimates, thus bolstering their confidence and willingness to participate in the next round of fundraising.
The major risks at this stage center on people and practices. You can hire people who don’t actually adopt the core value, or your onboarding, cultural, and technology practices don’t align or support this core value.
You will formalize and preserve this foundation by ensuring that the previous foundations remain strong; ensuring that your business and technology practices are data-driven; ensuring that everyone uses the KMS; and that key result data are democratized, and those data and metadata are high-quality.
Stage: Rapid Growth
Foundation Objectives: Operational resilience, scalability, IT operating model. 100% QA automation; high tech wrangler satisfaction; high org satisfaction with tech wranglers, cost control.
Key Results: Operational NFRs (MTTD, MTTR), QA and code metrics, cost/story point not increasing; increasing story points/sprint/person.
Tools: MCB, surveys, performance reviews, FinOps, team-chosen continuous improvement (Kaizen) programs.
Core value: If we get knocked down, we know it immediately and we get up again quickly (cuz we’re objectively awesome).
If you’re as successful as you hope, as you’ve worked so hard to be, then rapid growth exposes you to the highest probability of serious risk. Not necessarily of existential risk, but of significant damage to customers, your reputation, revenue, investor confidence, and possibly involvement of law enforcement or regulators, and, as always, being sued.
Operational resilience (OR) is your primary means of mitigating these risks. OR is the ability of an organization to prevent, adapt, respond, recover, and learn from operational disruptions, whatever their source. OR requires that the previous foundations have delivered on fundamentals like detection and recovery. Impact to business is your guide to prioritizing which components of your business will need the most work to establish a baseline of resilience.
The U.K.’s Financial Conduct Authority (FCA) has a set of simple steps that can help you understand how to achieve operational resilience.
For your customers, the foundation objectives of operational resilience, scalability, and 100% QA automation deliver the kind of product partnership that they can rely on to build their business. Your operational resilience becomes, in part, theirs. Your scalability allows them to scale, and your high-quality product helps them set their own high standard.
At this stage, investors get your solid track record of growth without floundering, giving them to confidence to invest at higher and higher valuations, hoping to stay in long enough for a big fat IPO.
The major risks to this value are the usual suspects, not prioritizing core values, hiring the wrong people, poor onboarding, and failing to adequately define and monitor OKRs for departments and team members.
You will formalize and preserve this foundation by:
a) hiring and empowering someone to handle OR/BC/DR, or else, and forgive me for saying this, forming a committee to do it, and then doing the work of planning, gaming, and learning;
b) hiring team members with demonstrated experience building scalable teams, and systems;
c) directing HR to craft a serious program of OKR definitions that should form a key part of dept./team performance reviews. Keep in mind that many managers have a very hard time coming up with objective measures for key results. Remember, nobody can exceed binary measures; subjective measures breed political behavior.
Questions, anyone?
Of course you have questions. Why did I say x? How do you do y? But what if we’re z? The next two articles in this series will answer many of your questions, but not all. Sorry, sorry, sorry. Kevin is the worst.
In your heroic journey to conquer the odds, there are one hundred and one pain points, and that’s just the first year. Many of them can be managed adequately by maintaining core values and priorities, and by building and maintaining the foundations above. For the ten most challenging pain points, I have some guidance, hopefully less dry than this first article:
Article 2
- No one tech person can do it all for the next five years
- You need different kinds of key people and skill-sets at each stage
- You need to do a better job of building the machine and pulling the levers
- Stop paying programmers to write code
- Why is this borken again?!?
Article 3
- Programmers and their dark secrets
- Data is your oxygen; don’t asphyxiate your business
- Other People’s Software Sucks
- Complexity is Startup Enemy #1
- AI isn’t a thing yet, but…
Finally, I’ll finish the third article with some advice on who to hire as soon as you can. Me! jk I’m the best there is, but I’m not available. Maybe…
In Conclusion
At the top of the article, I showed how easy and costly it can be to get this stuff wrong, whether you’re a startup or a massive multinational. Some of the blame can be laid on the MCB, and some on entropy within organizations, but entropy is not a mysterious force. Entropy happens when we don’t devote sufficient time to maintenance.
As granny would say, a place for everything, and everything in its place. Or as I would say, make using the KMS a core value and top five priority. Experience shows it’s easier to start it right and keep it right than it is to fix it two years down the road. Just like it’s easier to wash a dish now than when the Sriracha has dried to a substance harder than tungsten-carbide drill bits.
Good luck!
~Kev
OMG! Founder — Foundation. That just hit me!
