What Is Technical Debt? A Complete Guide

You buy the latest iPhone on credit. Turn to fast car loan services to get yourself those wheels you’ve been eyeing for a while. Take out a mortgage to realise your dream of being a homeowner. Regardless of the motive, the common denominator is going into financial debt to achieve something today, and pay it off in future, with interest. The final cost will be higher than the loan value that you took out in the first place. However, debt is not limited to the financial world.

Technical Debt Definition

Technical debt – which is also referred to as code debt, design debt or tech debt – is the result of the development team taking shortcuts in the code to release a product today, which will need to be fixed later on. The quality of the code takes a backseat to issues like market forces, such as when there’s pressure to get a product out there to beat a deadline, front-run the competition, or even calm jittery consumers. Creating perfect code would take time, so the team opts for a compromised version, which they will come back later to resolve. It’s basically using a speedy temporary fix instead of waiting for a more comprehensive solution whose development would be slower.

How rampant is it? 25% of the development time in large software organisations is actually spent dealing with tech debt, according to a multiple case study of 15 organizations. “Large” here means organizations with over 250 employees. It is estimated that global technical debt will cost companies $4 trillion by 2024.

Is there interest on technical debt?

When you take out a mortgage or service a car loan, the longer that it takes to clear it the higher the interest will be. A similar case applies to technical debt. In the rush to release the software, it comes with problems like bugs in the code, incompatibility with some applications that would need it, absent documentation, and other issues that pop up over time. This will affect the usability of the product, slow down operations – and even grind systems to a halt, costing your business. Here’s the catch: just like the financial loan, the longer that one takes before resolving the issues with rushed software, the greater the problems will pile up, and more it will take to rectify and implement changes. This additional rework that will be required in future is the interest on the technical debt.

Reasons For Getting Into Technical Debt

In the financial world, there are good and bad reasons for getting into debt. Taking a loan to boost your business cashflow or buy that piece of land where you will build your home – these are understandable. Buying an expensive umbrella on credit because ‘it will go with your outfit‘ won’t win you an award for prudent financial management. This also applies to technical debt.

There are situations where product delivery takes precedence over having completely clean code, such as for start-ups that need their operations to keep running for the brand to remain relevant, a fintech app that consumers rely on daily, or situations where user feedback is needed for modifications to be made to the software early. On the other hand, incurring technical debt because the design team chooses to focus on other products that are more interesting, thus neglecting the software and only releasing a “just-usable” version will be a bad reason.

Some of the common reasons for technical debt include:

  • Inadequate project definition at the start – Where failing to accurately define product requirements up-front leads to software development that will need to be reworked later
  • Business pressure – Here the business is under pressure to release a product, such as an app or upgrade quickly before the required changes to the code are completed.
  • Lacking a test suite – Without the environment to exhaustively check for bugs and apply fixes before the public release of a product, more resources will be required later to resolve them as they arise.
  • Poor collaboration – From inadequate communication amongst the different product development teams and across the business hierarchy, to junior developers not being mentored properly, these will contribute to technical debt with the products that are released.
  • Lack of documentation – Have you launched code without its supporting documentation? This is a debt that will need to be fulfilled.
  • Parallel development – This is seen when working on different sections of a product in isolation which will, later on, need to be merged into a single source. The greater the extent of modification on an individual branch – especially when it affects its compatibility with the rest of the code, the higher the technical debt.
  • Skipping industrial standards – If you fail to adhere to industry-standard features and technologies when developing the product, there will be technical debt because you will eventually need to rework the product to align with them for it to continue being relevant.
  • Last-minute product changes – Incorporating changes that hadn’t been planned for just before its release will affect the future development of the product due to the checks, documentation and modifications that will be required later on

Types of Technical Debt

There are various types of technical debt, and this will largely depend on how you look at it.

  • Intentional technical debt – which is the debt that is consciously taken on as a strategy in the business operations.
  • Unintentional technical debt – where the debt is non-strategic, usually the consequences of a poor job being done.

This is further expounded in the Technical Debt Quadrant” put forth by Martin Fowler, which attempts to categorise it based on the context and intent:

Technical Debt Quadrant

Source: MartinFowler.com

Final thoughts

Technical debt is common, and not inherently bad. Just like financial debt, it will depend on the purpose that it has been taken up, and plans to clear it. Start-ups battling with pressure to launch their products and get ahead, software companies that have cut-throat competition to deliver fast – development teams usually find themselves having to take on technical debt instead of waiting to launch the products later. In fact, nearly all of the software products in use today have some sort of technical debt.

But no one likes being in debt. Actually, technical staff often find themselves clashing with business executives as they try to emphasise the implications involved when pushing for product launch before the code is completely ready. From a business perspective, it’s all about weighing the trade-offs, when factoring in aspects such as the aspects market situation, competition and consumer needs. So, is technical debt good or bad? It will depend on the context. Look at it this way: just like financial debt, it is not a problem as long as it is manageable. When you exceed your limits and allow the debt to spiral out of control, it can grind your operations to a halt, with the ripple effects cascading through your business.

 

Check our similar posts

2015 ESOS Guidelines Chapter 3 ? The ESOS Assessment

ESOS operates in tandem with the ISO 50001 (Energy Management) system that encourages continual improvement in the efficient use of energy. Any UK enterprise qualifying for ESOS that has current ISO 50001 certification on the compliance date by an approved body (and that covers the entire UK corporate group) may present this as evidence of having completed its ESOS assessment. It does however still require board-level certification, following which it must notify the Environment Agency accordingly.

The Alternate ESOS Route

In the absence of an ISO 50001 energy management certificate addressing comprehensive energy use, a qualifying UK enterprise must:

  1. Measure Total Energy Consumption in either kWh or energy spend in pounds sterling, and across the entire operation including buildings, industrial processes and transport.
  2. Identify Areas of Significant Energy Consumption that account for at least 90% of the total. The balance falls into a de minimis group that is officially too trivial to merit consideration.
  3. Consider Available Routes to Compliance. These could include ISO 500001 part-certification, display energy certificates, green deal assessments, ESOS compliant energy audits, self-audits and independent assessments
  4. Do an Internal Review to make sure that you have covered every area of significant consumption. This is an important strategic step to avoid the possibility of failing to comply completely.
  5. Appoint an Approved Lead Assessor who may be internal or external to your enterprise, but must have ESOS approval. This person confirms you have met all ESOS requirements (unless you have no de minimis exceptions).
  6. Obtain Internal Certification by one of more board-level directors. They must certify they are satisfied with the veracity of the reports. They must also confirm that the enterprise is compliant with the scheme.
  7. Notify the Environment Agency of Compliance within the deadline using the online notification system at snapsurveys.com as soon as the enterprise believes is fully compliant.
  8. Assemble your ESOS Evidential Pack and back it up in a safe place. Remember, it is your responsibility to provide proof of the above. Unearthing evidence a year later it not something to look forward to.

The ESOS assessment process is largely self-regulatory, although there are checks and balances in place including lead assessor and board-level certifications. As you work through what may seem to be a nuisance remember the primary objectives. These are saving money and reducing carbon emissions. Contact Ecovaro if we can assist in any way.

How DevOps oils the Value Chain

DevOps ? a clipped compound of development and operations – is a way of working whereby software developers are in a team with project beneficiaries. A client centred approach extends the project plan to include the life cycle of the product or service, for which the software is developed.

We can then no longer speak of a software project for say Joe?s Accounting App. The software has no intrinsic value of its own. It follows that the software engineers are building an accounting app product. This is a small, crucially important distinction, because they are no longer in a silo with different business interests.

To take the analogy further, the developers are no longer contractors possibly trying to stretch out the process. They are members of Joe?s accounting company, and they are just as keen to get to market fast as Joe is to start earning income. DevOps uses this synergy to achieve the overarching business goal.

A Brief Introduction to OpsDev

You can skip this section if you already read this article. If not then you need to know that DevOps is a culture, not a working method. The three ?members? are the software developers, the beneficiaries, and a quality control mechanism. The developers break their task into smaller chunks instead of releasing the code to quality control as a single batch. As a result, the review process happens contiguously along these simplified lines.

Code QC Test ? ? ?
? Code QC Test ? ?
? ? Code QC Test ?
? ? ? Code QC Test
Colour Key Developers Quality Control Beneficiary

This is a marked improvement over the previously cumbersome method below.

Write the Code ? Test the Code ? Use the Code
? Evaluate, Schedule for Next Review ?

Working quickly and releasing smaller amounts of code means the OpsDev team learns quickly from mistakes, and should come to product release ahead of any competitor using the older, more linear method. The shared method of working releases huge resources in terms of user experience and in-line QC practices. Instead of being in a silo working on its own, development finds it has a richer brief and more support from being ?on the same side of the organisation?.

The Key Role that Application Program Interfaces Play

Application Program Interfaces, or API?s for short, are building blocks for software applications. Using proprietary software-bridges speeds this process up. A good example would be the PayPal applications that we find on so many websites today. API?s are not just for commercial sites, and they can reduce costs and improve efficiency considerably.

The following diagram courtesy of TIBCO illustrates how second-party applications integrate with PayPal architecture via an API fa?ade.

Working quickly and releasing smaller amounts of code means the OpsDev team learns quickly from mistakes, and should come to product release ahead of any competitor using the older, more linear method. The shared method of working releases huge resources in terms of user experience and in-line QC practices. Instead of being in a silo working on its own, development finds it has a richer brief and more support from being ?on the same side of the organisation?.

imgd2.jpg

The DevOps Revolution Continues ?

We close with some important insights from an interview with Jim Stoneham. He was general manager of the Yahoo Communities business unit, at the time Flickr became a part. ?Flickr was a codebase,? Jim recalls, ?that evolved to operate at high scale over 7 years – and continuing to scale while adding and refining features was no small challenge. During this transition, it was a huge advantage that there was such an integrated dev and ops team?

The ?maturity model? as engineers refer to DevOps status currently, enables developers to learn faster, and deploy upgrades ahead of their competitors. This means the client reaches and exceeds break-even sooner. DevOps lubricates the value chain so companies add value to a product faster. One reason it worked so well with Flickr, was the immense trust between Dev and Ops, and that is a lesson we should learn.

?We transformed from a team of employees to a team of owners. When you move at that speed, and are looking at the numbers and the results daily, your investment level radically changes. This just can’t happen in teams that release quarterly, and it’s difficult even with monthly cycles.? (Jim Stoneham)

Contact Us

  • (+353)(0)1-443-3807 – IRL
  • (+44)(0)20-7193-9751 – UK
What Kanban can do for Call Centre Response Times

When a Toyota industrial engineer named Taiichi Ohno was investigating ways to optimise production material stocks in 1953, it struck him that supermarkets already had the key. Their customers purchased food and groceries on a just-in-time basis, because they trusted continuity of supply. This enabled stores to predict demand, and ensure their suppliers kept the shelves full.

The Kanban system that Taiichi Ohno implemented included a labelling system. His Kanban tickets recorded details of the factory order, the delivery destination, and the process intended for the materials. Since then, Ohno?s system has helped in many other applications, especially where customer demand may be unpredictable.

Optimising Workflow in Call Centres
Optimising workflow in call centres involves aiming to have an agent pick up an incoming call within a few rings and deal with it effectively. Were this to be the case we would truly have a just-in-time business, in which operators arrived and left their stations according to customer demand. For this to be possible, we would need to standardise performance across the call centre team. Moving optimistically in that direction we would should do these three things:

  • Make our call centre operation nimble
  • Reduce the average time to handle calls
  • Decide an average time to answer callers

When we have done that, we are in a position to apply these norms to fluctuating call frequencies, and introduce ?kanbanned? call centre operators.

Making Call Centre Operations Nimble
The best place to start is to ask the operators and support staff what they think. Back in the 1960?s Robert Townsend of Avis Cars famously said, ?ask the people ? they know where the wheels are squeaking? and that is as true as ever.

  1. Begin by asking technical support about downtime frequencies, duration, and causes. Given the cost of labour and frustrated callers, we should have the fastest and most reliable telecoms and computer equipment we can find.
  1. Then invest in training and retraining operators, and making sure the pop-up screens are valuable, valid, and useful. They cannot do their job without this information, and it must be at least as tech-savvy as their average callers are.
  1. Finally, spruce up the call centre with more than a lick of paint to awaken a sense of enthusiasm and pride. Find time for occasional team builds and fun during breaks. Tele-operators have a difficult job. Make theirs fun!

Reducing Average Time to Handle Calls
Average length of contact is probably our most important metric. We should beware of shortening this at the cost of quality of interaction. To calculate it, use this formula:

Total Work Time + Total Hold Time + Total Post Call Time

Divided By

Total Calls Handled in that Period

Share recordings of great calls that highlight how your best operators work. Encourage role-play during training sessions so people learn by doing. Publish your average call-handling time statistics. Encourage individual operators to track how they are doing against these numbers. Make sure your customer information is up to date. While they must confirm core data, limit this so your operators can get down to their job sooner.

Decide a Target Time to Answer Calls
You should know what is possible in a matter of a few weeks. Do not attempt to go too tight on this one. It is better to build in say 10% slack that you can always trim in future. Once you have decided this, you can implement your Kanban system.

Introducing Kanban in Your Call Centre Operation
Monitor your rate of incoming calls through your contact centre, and adjust your operator-demand metric on an ongoing basis. Use this to calculate your over / under demand factor. Every operator should know the value on this Kanban ticket. It will tell them whether to speed up a little, or slow down a bit so they deliver the effort the call rate demands. It will also advise the supervisor when to call up reserves.

Contact Us

  • (+353)(0)1-443-3807 – IRL
  • (+44)(0)20-7193-9751 – UK

Ready to work with Denizon?

Contact Us Today