An Introduction to Agile Development, Part 2: Getting Started with Scrum

TL;DR: To get started with Scrum, grab one of the popular tools (e.g., Jira), implement all the Scrum artifacts and fill your gaps with some experienced people. It is not all too important to have the right task estimates, sprint lengths, etc. from the beginning or to follow the methodology dogmatically. The crucial part is to have a team that commits to the process, reflects critically, is eager to learn, and improves incrementally.

It took us a little while to get the second part of our series about agile development done, but Andreas and I had some pretty good reasons for it (or at least we hope so). While I moved back to Germany from the U.S. and started a new position as a UX Manager, Andreas is now not only the Head of Research at DATEV, but also started as a full professor of Web Engineering at Anhalt University of Applied Sciences.

But anyway, we somehow managed to also talk a little bit about Scrum. In this part of the series, we want to specifically focus on how to get started with it. There are many tutorials and introductions describing Scrum, such as Atlassian’s brief introduction to the topic. However, some questions that both Andreas and I have encountered over and over are often poorly addressed or not addressed at all. Let’s start at the very beginning:

Max: How do you get started with Scrum, i.e., when you want to implement it from scratch?

Andreas: When starting from scratch, in my opinion you should take one of the usual Scrum-supporting tool (like Jira, Meistertask, Sprintly, or Team Foundation Server) and begin by implementing the basic Scrum artifacts. That is, establishing the Scrum roles, establishing a backlog, the meetings (or “ceremonies”, in Jira’s words) etc. Based on my experience, it is important to follow the Scrum methodology at the beginning of the introductory phase closely, but not dogmatically. However, if the budget is available, then fill the gaps in your organization now (e.g., an experienced Scrum master). Additionally, I suggest to manifest the Scrum methodology explicitly by establishing a social contract in the team as a starting point of the transition process and to get the team members’ commitment. The social contract should cover the characteristics of the team, so that they know how to behave within the Scrum process, e.g., how to estimate task efforts (Scrum poker would be one possibility). Particularly, the Scrum master is responsible for permanently training the team and refocusing the process if required.

Max: I’ve experienced situations in which the Scrum master and the product owner were the same person. Do you think this is generally a bad idea?

Andreas: It is imperative that the two roles can be executed well with respect to an impact-driven product (the product owner’s responsibility) as well as an efficient process and a happy team (the Scrum master’s responsibilities). One has to be aware of the fact that often, these goals are competing, which requires a team setting where neither the product goals nor the team values and principles are sacrificed. Of course, it is hard to ensure a balance between these goals if the jobs are done by the same person. While it is the product owner’s job to get new features on the road as quickly as possible, the Scrum master has to defend the process and the scope of the sprint. However, similar problems might occur when appointing inexperienced people or one experienced and one less experienced person. My point here is that there are many situations in which an agile team might be influenced in a bad way, and not only because one person is in charge of the Scrum Master and product owner roles at the same time.

Hence, in my experience, the key to a successful agile team is a valid setting following agile principles. Appointing two experienced people capable of executing the roles as defined might be the perfect setting. Yet, one really experienced person doing the two jobs with some additional backup by the team can lead to success as well, but should not be a permanent situation.

Max: Once you’ve committed the team to the agile process, how do you determine the initial sprint length?

Andreas: For this, I suggest to first analyze the given goal. In many cases, Scrum is causing problems when the sprint length is not adjusted well. For instance, when new requirements are brought up in short iterations and thus break the sprint often, or because of a team not yet capable of applying the vertical slicing methodology properly, which often leads to unfinished sprints. Hence, start with a best guess, then observe the rough lenght of the cycles in which new requirements appear and adjust the sprint length accordingly.

Max: You’ve also mentioned task efforts before. Many teams seem to struggle in this regard. Do you have any suggestions for them?

Andreas: Very true! I’ve also observed that teams often struggle with defining some kind of useful estimation. I suggest starting with T-shirt sizes. Often, a best practice is that, together with the team, the Scrum master defines the minimum size as “S” (e.g., 1 day), then “M” (e.g., double of “S”), “L”, and “XL”. After some time and experience, the team will recognize that this approach might not be suitable for their complex tasks and look more into, e.g., Fibonacci-based story points and other more advanced complexity measures. From time to time, the Scrum master needs to carefully adjust, thus leading the team to a method which is appropriate, efficient, and—last but not least—accepted by the management. From this point of view, an objective suggestion might not be possible, since this is strongly dependent on the team setting and product situation. However, a common mistake is to have no agile-experienced person(s) in place. More than once, I observed teams failing to improve (in reasonable time spans) due to a missing experienced “guide” leading the team towards a productive setting.

Max: Following up on this, in which cases and under which circumstances would it be justifiable to work without task estimates?

Andreas: Well, there can, of course, be situations in which effort estimations are too difficult, inefficient, or simply frustrating the team. For instance, if no resource planning is possible due to external reasons, synchronization with many other teams is necessary, autonomy of teams is not ensured in general, or, particularly, transparency and trust are not established within the team or company, then it is a waste of time working on task estimates, since they will always fail. Hence, to ensure that the team is adopting the agile mindset, it is eligible to pause the estimation of tasks, in my opinion.

However, I strongly disagree with the sometimes expressed opinion that task estimates are not worthwile to work on or to improve, respectively. For an ambitious team, it is crucial to be self-aware regarding the impact of their work and the challenges (e.g., with respect to knowledge transfer) and to be able to justify particular tasks (see, for example, our discussion about technical goals in the first part of the interview).

Max: Getting back to the role of the Scrum master: Often, they are recruited from people who were not active in software development. What does a Scrum master who is (or was) not a software engineer themselves have to pay specific attention to?

Andreas: Well, again, this is a controversial issue. I think we can agree that software development is a very specific field driven by strong individual skills and team efforts. People with a background in software development mostly have experienced the specifics of this field, are prepared for upcoming situations, and simply speak the language of software engineers. Hence, they might be accepted easier by the team members, which is crucial for a Scrum master due to the lateral leadership relation in Scrum teams …

Max: … but this does not exclude people who were not active as software engineers?

Andreas: No, absolutely not! Actually, it is sometimes better to have a Scrum master who might team up with the product owner and help steer the process more towards product-related impact rather than focusing mostly on the technology, which happens, not always, but sometimes. However, with no background in software development one has to be really careful about common mistakes. For instance, with making statements about difficult tasks and assessing them as easy (or the other way around), the Scrum master might degrade himself in the eyes of the software engineers, who are usually very proud of their skills and their work.

Therefore, my tip for Scrum masters without a software development background, who plan to work with software engineers, is to observe a software development team for several sprints without taking a particular role and to learn the language, the specifics of the field and maybe even try to do do some tasks by themselves to experience the complexity of computing in general. This way, hopefully, the foundations can be laid for earning trust and being accepted as a valid contributor to the team goals.

Max: In comparison: What do Scrum masters who are software engineers themselves have to pay specific attention to?

Andreas: In my opinion, former software engineers have a particular advantage, if and only if, they embrace their role as a Scrum master and use their software development experience for tackling the challenges of their team. In many cases, I’ve experienced that a former software engineer had fewer problems of steering the process towards iterative delivery. However, the risk of micro-discussions is high, since software engineers often have a dedication for technical issues even if they are not responsible for implementing them themselves. But if—for example—retrospectives are wasted with technical discussions, the team will not evolve with respect to the process quality.

Hence, Scrum masters having worked as software engineers should really pay attention to the main characteristics of the Scrum master role: be focused on the agile mindset, improve the working environment, identify and clear obstacles, etc. Particularly, the relationship with the product owner needs to be carefully developed to establish the necessary positive atmosphere in the team. As product owners are often (unintentionally) considered as outsiders by the software engineers, it is very important to mediate this conflict by leaving some of the software engineering background behind.

Max: Is there a threshold in terms of the number of people below which applying Scrum doesn’t make sense?

Andreas: The creation of and working with Scrum artifacts requires some time. In my opinion, you need at least five people in a team, but if the Scrum master and product owner are exclusively assigned to one team, then more software engineers might be required, so that the delivery speed is high enough and the product owner and Scrum master are actually “utilized”. Additionally, it depends on the maturity of the team members and the company. Particularly, if the culture of the company is not following the agile principles, even proven scalable agile frameworks (like SAFe or LESS) have a high risk to fail. In such a non-agile environment I would suggest having a core team focusing on the demanded functionality and additional supporters at the intersection to other teams focusing on solving the obstacles.

In general, I would like to turn the question around: Do you need a Scrum process for, e.g., organizing just 3 software engineers efficiently? My pragmatic solution would be to ensure that they are working together using direct communication. If that doesn’t work, I suggest to temporarily hire an Agile coach and help them find a suitable way of collaborating—but without aiming directly towards a Scrum process.

Max: Can Scrum work without retrospective meetings? If yes, under which circumstances?

Andreas: A core principle of the agile methodology is to adapt and to improve. Retrospective meetings are the corresponding expression in Scrum language. Additionally, retrospectives, e.g., establish an official timespan for controlled experiments that might lead to improvements, and many more positive aspects. So, on the one hand, retrospective meetings are very useful—if not necessary—for successfully implementing Scrum. On the other hand, if there is a team in which the team members are trusting each other, are experimental, challenge each other during implementation, give permanent honest feedback, are transparent to outsiders, etc., then you don’t necessarily have to establish a retrospective meeting at the end of each sprint. In short: If the team is very mature and permanently working towards improvement, then the time span between retrospective meetings can be increased. Yet, skipping them altogether is not such a good idea, from my experience. There is always something that needs discussion and is difficult to resolve during daily work.

***

This concludes our deep-dive into the specifics of Scrum, in particular, the roles of the Scrum master and product owner, task estimates, and retrospectives. The key takeaway here is probably that it is not so important to adhere to a strictly predefined process. Instead, aiming for a positive environment—with a team and leaders that are able and willing to adapt and continuously strive for improvement—will lead to efficient and happy teams.

In the third and final part of our interview, we will have a closer look at agile methodologies in the context of start-ups and the future of agile development in general. Thank you for reading and stay tuned!

Advertisements

An Introduction to Agile Development, Part 1: Scrum vs. Kanban

TL;DR: Two of the most popular agile methodologies are Scrum and Kanban. They mainly differ in the handling of prioritization (once per sprint vs. continuous) and deadlines (yes vs. no). Yet, independent of the approach, the key to success is to ensure focus and continuous delivery by: 1) creating small enough, manageable tasks, 2) prioritizing based on impact, and 3) finding a good trade-off between technical and business goals.

Many teams in software development and beyond are using agile processes to manage their tasks. However, working in an agile environment raises a variety of questions. Sometimes this is due to inexperienced team members, sometimes due to management deciding to use agile methods without having understood them.

Today, I have the pleasure to discuss agile development and two of the most popular agile methodologies—Scrum and Kanban—with Dr. Andreas Both, the Head of Architecture, Web Technologies, and IT Research at DATEV. Before, Andreas and I worked together at Unister, Germany’s biggest E-commerce company at that time, where he supervised my Ph.D. thesis. In this series of articles, our goal is to make agile development processes more understandable to people new to the topic and to dive a little deeper into the specifics of each methodology.

Max: To kick off the interview, we are going to start with one of the questions that is posed most often when people get started with agile methodologies and that I had to answer on way more than one occasion: What is the difference between Scrum and Kanban?

Andreas: In my honest opinion the core difference is the time taken for prioritization and iterations. While in Scrum the team is prioritizing once in an iteration, and during the sprint mostly only pull tasks, in Kanban the tasks are re-prioritized continuously. Hence, a task might be the next to be pulled from the backlog even though it has been raised by the product owner just seconds ago. Based on my experience, Scrum teams are feeling more freedom and less pressure due to the fact that in most companies the sprint is a protected area where external disruptions are minimized. Hence, development teams have more options to organize and plan their work, which might lead to higher happiness.

Max: Are there also differences in handling deadlines between Scrum and Kanban?

Andreas: A Scrum team delivers at the end of the sprint the latest. Kanban teams do not need an actual deadline, but if it takes very long to finish a task this might indicate a problem with slicing the task at hand into small enough, manageable subtasks. However, this problem also appears in Scrum teams. So, the differences with respect to deadlines are not significant if the process lead takes care. I think the core capability is to maintain a continuous delivery process.

Max: What is the best way or process to prioritize tasks?

Andreas: Based on my experience I’d say an impact-driven approach has the best chances of resulting in a successful project. It provides the opportunity for failing fast if the project is not executable. However, a mature team is needed in most cases to ensure success following this approach. For instance, the capability of slicing large tasks into smaller ones is required, so that manageable tasks are created and can be prioritized. Within inexperienced teams, you often hear statements like “We can’t slice it into smaller tasks,” which frequently leads to saving-the-world huge meta-tasks. Of course, there is no best process for task prioritization—just suitable ones for the current team’s or product owner’s maturity level and the product context. For instance, a simple approach might also help to establish a good acceptance of the product by the users. Finally, to live a methodology is crucial. A product owner should never be in the situation to not provide a justification or methodology for his prioritizations.

Max: So, from this, I conclude the process is not crucial in your opinion as long as the prioritization ensures focus. A different issue is that while establishing a backlog many teams struggle to prioritize requests of different kinds. How do you handle the conflicts between business goals and technical goals?

Andreas: I think there is no short answer to this question. When starting a project I suggest to favor the business needs over the long-term maintainability. This is due to a typical observation that requirements are unclear at this point in time. Early ideas about useful architecture etc. often do not hold and lead to wasted investments. At the same time, you have to communicate permanently to the (product) management that these early results and the velocity at which they have been created cannot be considered as regular and long-term. After some time teams tend to claim a certain amount of time to work on reducing technical debt. It is a question of maturity of the organization whether the development team can actually invest time to achieve technical goals or not. Unfortunately, we had to fight hard for this and lost many times.

Max: Isn’t there a risk that this leads to unsatisfied developers?

Andreas: Yes, there is a risk for sure. Unfortunately, reality shows that it is really hard to argue against killer arguments from strategic product management like “If we are not finishing this feature first, then the competitors will win.” Therefore, the Scrum master and the product owner have to work seriously on communicating the global priorities of feature completeness, maintainability, adaptability, etc.

Max: Very true. But what is your suggestion for a fresh product owner for prioritizing also non-business goals, such as the elimination of technical debt?

Andreas: In my experience, the key to success is to transform technical goals into business goals by finding related business KPIs. Let’s assume that the development team is annoyed because there is no time to refactor the source code and implementing unit tests for a certain component. While doing some analytics one might discover that the same component often contains bugs which need to be fixed in the sprint instantly. Hence, the missing tests are a reason for missing the deadlines for completing new features—leading to the business goal effectively encapsulating the technical goal. If the process lead and the product owner work closely together, then these justifications might be easy to find. However, if you do not find a reason, you should also talk frankly to the development team that it will not happen … and sometimes you just need to show courage and frankly inform your product management about the plan of aiming for a technical goal and just do it without justification.

Max: What are typical examples of features that should be realized using Scrum on the one hand and Kanban on the other?

Andreas: In my opinion, Scrum is more suitable for development teams targeting product goals where they can predict the requirements of the business domain (at least a little bit). If this is not the matter, then blockers within the Scrum sprint are more likely, which reduces the predictability of results at the end of a sprint. On the other hand, Kanban can be applied more easily with teams working in less predictable fields (e.g., research teams) or which are triggered by external interrupts (e.g, IT operations teams). For larger teams, there are results from different domains that Scrum is doing better. However, if you have such a large project that several teams are required, then your teams really should be experienced in agile methodology anyways.

This concludes the first part of Andreas’s and my little introduction to agile development, where we mainly addressed general issues such as deadlines, prioritization, communication, and the differences between business and technical goals. Also, we’ve had a closer look at Kanban and how it is different from Scrum. In the next part of the interview, we will be paying more attention to the specifics of the latter. Stay tuned!

What ’bout some fancy dashboards for ya? Power BI vs. Geckoboard

In my capacity as the chief data analyst of bitstars, it’s one of my key responsibilities to regularly compile all relevant figures concerning our web platform HoloBuilder. These figures are mostly intended for people who don’t have the time to dive deeply into some fancy but complicated statistics. Hence, from the user experience perspective it’s crucial to provide them in an easy-to-understand and pleasant-to-look-at form. A well-established way of doing so are data visualizations of different forms which are provided in terms of dashboards for optimal accessibility. Since we are currently redesigning our internal process for providing figures and statistics, I’ve done some research on two potential software solutions that could be used for this.

Requirements

Since we are talking about a solution for our internal process at bitstars, there is a set of company-specific requirements a contemplable software has to fulfill. In particular, these requirements are:

Moreover, there are two nice-to-haves:

In the following, I investigate two possible solutions—Power BI and Geckoboard—, which are evaluated against the above requirements.

Power BI

Power BI WebsitePower BI is a cloud-based business analytics service provided by Microsoft. It comes as a part of the Office 365 suite, but can also be used standalone. There is an online as well as a desktop version, whereas the latter has a significantly larger range of functions. Power BI distinguishes between dashboards (“[…] something you create or something a colleague creates and shares with you. It is a single canvas that contains one or more tiles.”) and reports [“one or more pages of visualizations (charts and graphs)”]. Reports can be saved in the Power BI Desktop file format (.pbix); dashboards can be shared.

Power BI comes with a rather limited range of integrable services, among which are Google Analytics (☑) and MailChimp (☑). AdWords (☑) statistics can be integrated via Google Analytics if your respective accounts are connected. However, integration for Facebook Ads (✖), Pipedrive (✖) and AWS (✖) is still missing. FB Ads integration has been requested, but is yet to be realized. There is moreover functionality to integrate data from Excel and CSV files (from your computer or OneDrive) or Azure SQL databases, among others, which also enables you to import your own custom data.

The basic version of Power BI can be used for free while Power BI Pro comes for $9.99 per user & month.

How to create a cumulative chart in Power BI?

Cumulative charts are not a built-in functionality of Power BI, but can be easily realized using Data Analysis Expressions (DAX, ☑). That is, you have to create a new measure in your dataset. Assume, for instance, you want a cumulative chart of your sales (to be accumulated, Y axis) over time, which are only present in your dataset as the number of sales per date (X axis). The DAX formula for your new measure would be as follows:

Measure = calculate(
  SUM('Your Dataset'[Sales]);
  FILTER(
    all('Your Dataset'[Date]);
    'Your Dataset'[Date] <= max('Your Dataset'[Date])
  )
)

(found at http://www.daxpatterns.com/cumulative-total/). You can then simply add a chart visualizing your new measure (Y axis) per date (X axis) to your Power BI report to obtain your desired cumulative chart.

Geckoboard

Geckoboard WebsiteGeckoboard is a web platform for creating individual dashboards that show your business’s KPIs (key performance indicators), e.g., unique visits to your website, Facebook likes or sales per day. The platform has built-in support for integration of a wide range of external data sources, including Google Analytics (☑), AdWords (☑), Facebook Ads (☑), MailChimp (☑), Pipedrive (☑) and AWS (☑) and many more (in fact, way more compared to Power BI). Moreover, Geckoboard supports CSV and Google Sheets integration for your own custom data.

Like in Power BI, there is no built-in support for cumulative charts. However, since it is easily possible to create those in Google Sheets (see, e.g., this link), they can simply be imported and visualized in Geckoboard as well (☑). Of course, this means an additional intermediate step is required.

Geckoboard offers no free plan. Paid plans start from $49 per month for one user and two dashboards.

Conclusion

Power BI Geckoboard
Cumulative charts (☑)1 (☑)1
Google Analytics integration
AdWords integration
Facebook Ads integration
MailChimp integration
(Pipedrive integration)
(AWS integration)
overall rating ⭐⭐⭐ ⭐⭐⭐⭐

Both tools miss built-in functionality for cumulative charts, but provide means for importing own custom data. When it comes to the integration of third-party services, Geckoboard supports a significantly lager range of available data sources. Because of this, I give Power BI an overall rating of 3 out of 5 (⭐⭐⭐). Since the pricing is more expensive and cumulative charts require an additional intermediate step, but the overall package makes a better impression regarding what we need at bitstars, Geckoboard receives a rating of 4 out of 5 (⭐⭐⭐⭐).

To summarize, if you’re fine with Google Analytics stats and some custom data imported via Excel files or an Azure DB, go for Power BI. Yet, if you rely on the seamless integration of a wider range of external services, you’re clearly better off with Geckoboard—unless you wanna implement the integration of the different services’ APIs yourself in a DIY solution.

1 These are given in parentheses because an additional intermediate step is required.