There is a longstanding joke in the Agile community that the real purpose of a Software Development Manager is to stay out of everyone’s way.

With self-governing teams interacting with competent Product Owners, an Agile Team can competently produce value in short increments, consistently moving the business further down the road to where it wants to be next. This team collaboration requires focus, discipline, communication, skill, competency, teamwork, and emotional intelligence by everyone involved.

Although humorous, this legitimately does raise the question—While the teams are busy generating value, what is the REAL purpose of the Development Manager?

The software development value stream has often been described as an assembly line of decisions, building a massive mathematical product. The better that humans can harmonize and organize, the better the product’s final utility value and fit will be for the organization’s customers.

Good Software Development is NOT Intuitive

All of this complexity must be managed correctly along the software value steam, and surprisingly it’s not intuitive. What you think you should do next is sometimes NOT what you should do next.  In other words, on-the-job training will produce pretty good managers, but not excellent ones.

The Standish Group’s CHAOS Surveys tell us what a typical organization look like.  Only 31% of projects are delivered successfully. 50% are late. And 19% are complete loses. (Standish 2020). With an combined industry budget of around 250 $Billion, this is a lot of unnecessary waste, dealt out by a lot of mediocre development leaders.  How do we fix this?

Ten Critical Software Development Bottlenecks:

There are ten critical bottlenecks (read: wastes of money) in the software development process and they are not intuitive. I propose the real job of a Software Development Manager is to attend to these things:

1. Employee Turnover – While this happens in our industry, every time a programmer leaves it takes about 6 months of work for the next one to come up to speed firing on all cylinders. Even highly skilled developers need time to acclimate to the products, the code base, the tools used in that environment, the customers, the customers attitudes, and how they use the product. Figuring a typical developer salary corporate burden today is around $150,000/annually, every time a programmer leaves a company, it costs them about $75,000 and six months of one person’s schedule.  Great Development Managers minimize turnover, saving the company lots of money and unnecessarily lost time.
   
   A good Development Manager keeps the teams productive and content.  Being near them, praising them for work well-done, and keeping the environment healthy and productive are important factors for healthy teams.  The best environments report high-trust, good communication, happy employees, and a good sense of satisfaction. Some people call this Psychology Safety.  
   
   It’s important to develop a culture of accomplishment and teamwork.  Team branding with a team identity, vision, mission, goals, and values which are supported by a feeling of appreciation and accomplishment is the magic formula here.
   
   

2. The Hidden Factory: I like to say “If it takes your team eight weeks to get six weeks’ worth of production completed, your team has 2 weeks of hidden factory.” This term describes everything that has to happen over again because it wasn’t done correctly or completely the first time around.  Meeting with customers again for the same reasons, re-structuring screens or reports, changing workflows, re-pushing builds, hot-fixes, testing churn, and many smaller and larger things including multi-million-dollar entire solution rejections which we see occasionally in large companies. These become multi-million-dollar mistakes.
   
   It’s not easy to address a Hidden Factory problem. Most groups naively think they just need more time to get some things done and if management would relax a bit they could catch up and stop making these consistent mistakes and omissions.  The truth is, they will never address the problem until they start doing work differently.  What they need is a quality system such as CMMI, ISO, or The Stable Framework to solve their reactivity issues.  This is an unintuitive reality of software development.
   
   Most groups I talk to report about a 33% effort loss to the Hidden Factory over time.  Just do some simple math on the total payroll burden of your Software Department x 33% and that’s only PART of what you are losing unnecessarily.  Add to that the opportunity cost for revenue delays on your products until 33% past the optimum endpoint and the wasted costs magnify.
   
   It is the Software Development Managers job to bring in a quality system to solve this problem. Quality pioneer W. Edwards Deming was extremely unpopular for putting the blame for low-quality work on management—not employees. Although he was uncomfortable and non-intuitive, he was right.
   
   

3. Not Using AI to Vibe Code: There is a tremendous advantage available now to software development teams who use AI to generate units of code and debug challenging problems. Many groups are reporting about a 30% increase in productivity using these tools. CoPilot, Clive, Grok.com, Gemeni, Claude.ai, and even local LLM’s using Ollama or LM-Studio will provide immediate lift.  To keep up with who’s currently leading the AI wars check out ARCPrize.org.
   
   
4. Using Awkward Tools: The first time I met Ken Schwaber, co-inventor of Scrum, he told me the best Scrum tools were a white board, sticky-notes, and a spreadsheet.  I met him again nine years later on the other side of the planet and he told me the same thing.
   
   Of course, today, many of us work remotely making those tools awkward. There are many digital tools to chose from. I can tell you from personal conversational experience with many professionals the most popular tools used tool today are Jira (80%), and SmartSheet (20%), with many companies using both.  Probably 10% of the remaining companies use Asana, Shrike, Monday.com, or ADO (Azure DevOps).
   Some companies are even reverting back to a virtual whiteboard and virtual sticky-notes using a group collaboration space such as Miro.com, LucidSpark, or just the new whiteboards inside of Zoom.  This works, and is a fun solution.
   
   Whatever tools your groups choose, make sure it helps keep you organized for speed, and doesn’t slow down your work. Incorrect tools, or misconfigured tools will slow the team down.
   
   
5. Bad Approaches to Coding: The first part of Robert C. Martin’s book Clean Architecture explains how important correct coding architectural techniques are to minimizing the cost of maintaining a developed system once deployed. Not all coders are the same, and if you hire sloppy coders you will soon have a production product that requires an exponential number of additional programmers just to maintain it.  Data with compelling examples are in his book. Basically, spaghetti code kills productivity.
   
   This costs a company millions of dollars of unnecessary payroll over time. As an owner, or investor, and because this is non-intuitive, you won’t know what you don’t know until it’s too late—and you will be paying a large team, instead of paying yourself.
   
   

6. Continuing to invest in Monoliths: Software development has evolved. In the past we all invested in Microsoft, Oracle, and other monoliths because they were the mainstream providers of proven technology. “Nobody ever got fired for buying IBM.” or so the expression went.
   
   While this once was an expensive but safe position, today Microsoft IIS webservers only make up about 4% of the internet.  Open-source solutions like PostgreSQL, Python, JavaScript, are proven, and all you need to construct a complete enterprise, or SAAS solution.  Forgive me for stating the obvious, but the days of these big players are waning.  Don’t spend unnecessarily big money on new investments here. There is no need anymore.
   
   
7. Hiring Wrong: An accountant and a seasoned software development manager would have two very different approaches to hiring for tech talent. While accountants want reasonable value for the best price, Steve Jobs said he witnessed a 25:1 difference in productivity between good and bad developers.
   
   You want your development staff to match the workload. Staffing for new development should look like an upside-down birthday cake.  You want it top heavy with senior coders, and just a few junior coders on the bottom to do the busy-work. This configuration will yield much more productivity over time of much and more stable products.
   
   It takes developers years to understand and permanently distill the larger compliment of common logical structures available within most programming languages.  Senior developers look at a challenge and select and organize which “tools” to use to construct the solution.  This is much like making a product from Legos.  Once you have a few years working with them, you’ll command a solid knowledge of all the common shapes.  Developers with this level of competency can be given an objective and can quietly create a solution correctly, the first time.  In contrast, junior developers must search and forage for solutions to what are all-new challenges for them.  This is the process that takes the x25 times longer Steve Jobs was talking about.
   
   
8. Not pulling weeds: Nobody likes a bad apple. Sometimes, despite all initial attempts for reaching across the aisle, you end up with a team member that causes problems for others on the team. This can quickly become a problem that never goes away without intervention. If someone presents with unwanted behavior over time, a good development manager should take the team member aside and find if the problem is temporary. If so, work with the team member privately to help them regain a successful gait. If not, they are likely in the wrong role. Work with them to get them where they would rather by. A good manager wants their team members to be successful not just on the job, but also in life. Good managers understand their team members are also their customers.
   
   
9. Outsourcing Incorrectly: Outsourcing is rarely effective. It’s never really been a good idea. Sure, you can find senior level management who defend outsourcing, but talk to any team and they will tell you it’s awkward, comparatively slow, and frustrating.  Most companies outsource because other companies outsource.  The argument used to be it was cheaper labor, but those days are past. Now the argument is the resources are instantly available.  This is a pretty weak argument. In my experience most senior executives support outsourcing so they can confirm to their senior board of directors, who are even more removed from the real software development value stream, that they are in-fact outsourcing, too.
   
   Companies who outsource successfully do it a certain way.  They budget for one member of their team to go stay with the outsourcing group for the duration of the project, or at least, go back and forth frequently.  Sometimes the model is reversed and the outsourcing agency has a team member stateside.  That team member spends 9 hours working with the American team, and then the other 9 hours working with the offshore team—each day.  It is brutal, and frankly unfairr
   
   Studies have shown outsourcing adds 20-40% on to a projects effort in cost and schedule—and that’s if you have a useful product.  Many a local team overcompensates for the outsourced contribution. However popular, it is difficult to outsource efficiently.
   
   
10. Not Investing in Training: Niccolo Machiavelli stated “…training--not budgets--wins wars!” That’s an important principle to digest.  More recently, Deming taught his customers, when evaluating vendors, to ask them how much money they had spent on training their people during the past 12 months.  This will give you some insight into the level of performance you’ll get out of them.
    
    

    Deming would draw the effect of training as a tight squiggly line representing effort, which abruptly straightens into a longer mildly jagged line at the point of training.  In this manner he demonstrated how training narrows the amplitude of ineffective effort, transforming the trainees’ efforts into work that move the company down the road towards their objectives faster.
    
    Training helps both new employees and seasoned employees get better at their jobs.  While new employees benefit from structured guidance on how to perform their tasks, seasoned employees benefit from having real-world experience they can conceptualize into immediate practical application.
    
                 
    
    Training also gives trainees a correct vocabulary use throughout the industry, and gives them confidence their efforts and skills are industry-standard.

In summary, statistics show software development is not performed efficently, and these ten factors are keys to leading your teams to perform much better than the Standish CHAOS Report data--or industry average of on 31% of project's being delivered on-time.