0:00:00.000,0:00:03.600
Thank you for the introduction. In the next 10 minutes I'm going

0:00:03.600,0:00:07.080
to give you a peek into the field of software effort estimation

0:00:07.080,0:00:11.820
and I'm going to show you how you can tackle the delays in your software deliveries.

0:00:11.820,0:00:19.800
I will also discuss why it's important to take a holistic approach to on-time software delivery.

0:00:21.300,0:00:25.800
So we all know that late delivery and cost overruns have been common problems

0:00:25.800,0:00:32.400
in the software industry for decades. On average software projects run over

0:00:32.400,0:00:38.820
time around 30% and this percentage does not seem to have decreased since the 1980s.

0:00:39.540,0:00:43.800
And even though effort estimation is at the heart of almost all industries,

0:00:43.800,0:00:47.640
it's especially challenging for the software industry

0:00:47.640,0:00:53.520
because software development is a complex undertaking that is affected by a variety

0:00:53.520,0:00:59.400
of vast social and technical factors. While currently most software companies

0:00:59.400,0:01:06.240
they rely on human experts' subjective assessments to arrive at an effort estimate.

0:01:06.240,0:01:12.540
And human experts use their intuition, they use rule of thumb, or they use generalized models.

0:01:12.540,0:01:19.020
And research has shown that this may lead to inaccurate and inconsistent estimates

0:01:19.020,0:01:24.000
because software development is, yeah, there are too many unknowns and too many

0:01:24.000,0:01:28.980
interrelated factors to reduce it to a gut feeling or to a rule of thumb.

0:01:29.640,0:01:34.440
So we need to have a deeper understanding of the factors that play a role here,

0:01:34.440,0:01:38.700
how they interact with each other, and we need to incorporate them into

0:01:38.700,0:01:42.960
our effort estimation models to reduce the delay in software deliveries.

0:01:45.000,0:01:51.000
So to investigate the most relevant factors and their interactions we performed case

0:01:51.000,0:01:53.700
studies at ING. ING is a large

0:01:53.700,0:01:58.980
Dutch internationally operating bank. And we use the mixed methods approach - so

0:01:58.980,0:02:04.620
we combined qualitative survey data with quantitative repository data

0:02:05.520,0:02:13.620
to to corroborate the respondents' perceptions and to also validate - to validate their perceptions

0:02:13.620,0:02:19.320
and to capture multiple perspectives on the phenomenon of on-time software delivery.

0:02:20.100,0:02:25.620
And we wanted to find out what factors affect delays in software delivery and

0:02:25.620,0:02:27.540
how are these factors related to each other.

0:02:29.760,0:02:38.760
And with such insights we can actually better understands what data and techniques are needed

0:02:38.760,0:02:43.920
to improve effort estimation methods and to predict and manage delay risks.

0:02:45.600,0:02:53.160
Okay, so from the open-ended survey responses and to - our validation with the repository data

0:02:53.160,0:02:59.100
we identified 25 factors that affect delays in software deliveries and we

0:02:59.100,0:03:06.840
organize these factors along five risk dimensions. So the first one is the organizational dimension

0:03:06.840,0:03:12.600
which actually concerns the uncertainty surrounding the organizational environments

0:03:12.600,0:03:18.600
in which your software delivery takes place. And three of the organizational factors are ranked

0:03:18.600,0:03:25.380
to be among the top five most influential factors. So, for example, organizational alignment, which

0:03:25.380,0:03:32.580
is actually having a shared vision and mission, which can ensure the alignments between

0:03:32.580,0:03:36.360
the implementation of your software delivery and your business strategy,

0:03:36.360,0:03:40.740
and this is reported to contribute to your timely software delivery.

0:03:41.700,0:03:49.020
Respondents also reported that they are often hampered by organizational politics

0:03:49.020,0:03:52.620
and the geographic distribution of the software teams.

0:03:54.120,0:03:58.320
Next we have process factors which represents the

0:03:58.320,0:04:03.720
effectiveness and the maturity of the way your software teams work.

0:04:03.720,0:04:09.408
And in this category, requirements refinements is the top mentioned factor that

0:04:10.500,0:04:15.660
affects the delay in your software deliveries. In fact it's the process of defining your

0:04:15.660,0:04:19.560
software delivery and breaking it down into smaller work items,

0:04:19.560,0:04:27.240
and missing or lacking details is reported to be the main delay factor in software deliveries.

0:04:29.100,0:04:33.840
We also have project factors which represents the inherent complexity

0:04:33.840,0:04:39.180
and uncertainty of a software project. And in this category, task dependencies

0:04:39.180,0:04:46.260
constitute the top mentioned delay factor. And these depends dependencies can occur across

0:04:46.260,0:04:53.580
teams and across products due to analyzed priorities and inconsistent schedules.

0:04:56.340,0:05:02.100
We have the people related factors as well. So respondents reported that teams that are

0:05:02.100,0:05:05.820
more stable, that are more skilled, more familiar with each other,

0:05:05.820,0:05:10.260
they they more often deliver software on time.

0:05:10.800,0:05:18.000
And the final category are the technical factors, and respondents reported to be often delayed by

0:05:18.540,0:05:22.440
technical dependencies and issues related to poor code documentation

0:05:22.440,0:05:27.180
and unreliable IT infrastructure. But the factors in this category,

0:05:28.140,0:05:33.300
they are reported to have the smallest impacts across all categories.

0:05:34.680,0:05:37.200
So to summarize, there's an overview

0:05:37.200,0:05:42.840
of the top five factors that have the greatest impact on the delays in your software deliveries.

0:05:43.980,0:05:48.000
So these are the factors that you should pay a great deal of attention to.

0:05:48.600,0:05:54.000
And the interesting thing is that none of them are in software at all, right,

0:05:54.000,0:05:58.980
they are related to your organizational environments and the quality of your processes,

0:05:58.980,0:06:04.140
so how your software teams work. And we can also see that

0:06:05.700,0:06:10.440
they are also controllable. So these are the factors that

0:06:10.440,0:06:14.460
software teams often complain about, right, this is something that is happening to us,

0:06:14.460,0:06:19.020
we don't have any impacts on it, but you can act upon upon

0:06:19.020,0:06:22.980
your requirements refinements, you can invest in your dependency management,

0:06:22.980,0:06:27.240
and you can also invest in a more healthy organizational culture.

0:06:30.600,0:06:36.900
So we know that these are the top factors that have an impact on your delay,

0:06:37.920,0:06:41.280
so this means we should regulate them, but then the next question is

0:06:41.280,0:06:44.760
how should we regulate them? So to answer that question we looked

0:06:44.760,0:06:49.200
into different types of relationships between factors and on-time delivery.

0:06:50.460,0:06:54.060
We more specifically focused on three types of orthogonal effects.

0:06:54.060,0:06:59.520
So we looked into direct relationships, indirect relationships, and contributory relationships.

0:07:00.960,0:07:04.980
And this enabled us to make a distinction between simple, complex,

0:07:04.980,0:07:10.320
and condition-dependent relationships. I'm not going into the very specific

0:07:10.980,0:07:14.820
interactions between factors but this actually enables

0:07:14.820,0:07:19.560
us to put all the results together. So the factors - the factor rankings

0:07:19.560,0:07:25.980
and their relationships to come - to create a conceptual framework of on-time software delivery

0:07:25.980,0:07:30.180
as shown here. So it might look a bit overwhelming at first,

0:07:31.560,0:07:34.800
but it actually visualizes the results that we discussed so far.

0:07:34.800,0:07:39.300
So you can see the factors in the ellipses, the categories in the blue rectangles,

0:07:39.300,0:07:44.700
and the arrows visualize the sign and the direction of the relationships that we found.

0:07:45.420,0:07:51.540
And the interesting thing is that factors interact hierarchically with each other.

0:07:51.540,0:07:55.080
So organizational factors interact with people factors,

0:07:55.080,0:07:57.840
which in turn impacts the technical factors,

0:07:59.340,0:08:06.420
and these hierarchical interactions indicate that changing one factor may impact another.

0:08:06.420,0:08:12.900
This shows that on-time software delivery is affected by a system of interrelated factors

0:08:14.700,0:08:20.340
that you should look at as a whole. So you should use a holistic approach

0:08:20.340,0:08:25.440
to tackle the delays in your on-time software - in your software delivery.

0:08:25.440,0:08:31.740
So don't try to tackle the factors individually but actually formulate systematic treatments

0:08:31.740,0:08:37.320
based on these direct and indirect relationships to have the greatest impact on your delay.

0:08:37.320,0:08:42.480
So I'll give you a concrete example. So the organizational factor "executive

0:08:42.480,0:08:48.900
support" is perceived to have an indirect impact on timely software delivery through team stability

0:08:50.100,0:08:55.440
and team stability directly affects how effective a team is in building

0:08:55.440,0:09:00.000
skills and knowledge over time and this affects the technical

0:09:00.000,0:09:06.840
factors code quality and bugs or incidents which introduce necessary rework for the

0:09:06.840,0:09:13.560
software team and thereby delay. So this means that establishing a

0:09:13.560,0:09:18.420
stronger executive support may lead to more stable - more skilled teams

0:09:18.420,0:09:25.740
that are better able to - to deal with - to maintain code quality and that can better resolve

0:09:25.740,0:09:31.500
delays that are caused by bugs or incidents. And this path from higher executive support

0:09:31.500,0:09:36.300
to the technical factors is an example of many similar paths

0:09:36.300,0:09:40.980
that we can derive from our conceptual framework and that enable us to

0:09:41.820,0:09:48.300
yeah make hypotheses about what can actually improve your on-time software delivery.

0:09:49.260,0:09:54.840
So what have you learned from this? Well first we saw that on time software

0:09:54.840,0:10:01.140
delivery is affected by a system of interrelated factors that require a holistic approach.

0:10:03.300,0:10:08.220
Secondly we saw that the top most influential factors are not in software they are actually

0:10:08.220,0:10:12.060
in your organizational environments and in the quality of your processes

0:10:12.060,0:10:18.300
so make sure to invest in your requirements refinements, effective backlog performance,

0:10:18.300,0:10:23.040
a better dependency management, and in a healthy organizational environment.

0:10:25.020,0:10:30.120
And at the same time, make sure to incorporate these factors in your effort estimation models

0:10:30.120,0:10:34.980
because when you capture information not only about the software task but also

0:10:34.980,0:10:37.740
about the processes, your team, your organization,

0:10:37.740,0:10:43.500
you can actually enhance the the predictive power of your effort estimation models.

0:10:46.080,0:10:51.000
Yes, so if you're interested in this work and if you would like to know more

0:10:51.000,0:10:54.060
about the conceptual framework, how to apply it in practice,

0:10:54.060,0:10:59.220
then feel free to contact me and thank you. All all,

0:10:59.220,0:11:03.720
right thank you very much Elvan, questions coming in from our viewers,

0:11:03.720,0:11:12.840
and the the first one is, where and how can I learn these quote political skills as a programmer

0:11:12.840,0:11:18.240
because they certainly weren't part of my degree. Well none of mine as well,

0:11:19.800,0:11:20.520
yeah,

0:11:20.520,0:11:25.200
based on my own experience it's something that you learn over time I would say,

0:11:25.200,0:11:36.960
so I - I would say that's - if you - if you end up in more situations where you actually need these

0:11:36.960,0:11:39.420
political skills to, for example,

0:11:40.200,0:11:45.660
defend your products, to actually create business strategy,

0:11:45.660,0:11:49.560
I think it will come with with experience but also maybe with age.

0:11:51.780,0:11:59.460
And I really - I think this is a related question. Is this the role or is - sorry I'm

0:11:59.460,0:12:04.680
just trying to make sense of this - is this something that product managers or project

0:12:04.680,0:12:12.180
managers are better able to do than developers? Is that part of their role - to handle these

0:12:12.180,0:12:14.700
non-technical issues? Yeah,

0:12:14.700,0:12:20.700
that is definitely part of their responsibilities. So in my research I've mainly looked at agile

0:12:20.700,0:12:24.600
projects where they work with, hey, you have a software development team

0:12:24.600,0:12:32.160
and you have a product owner that's the person who actually represents the

0:12:32.160,0:12:37.860
interest of the the client - the customer - and it's also the person who bridges the gap

0:12:37.860,0:12:43.740
between the software developers and the managers. But I think that would actually be nicer if we

0:12:43.740,0:12:47.160
wouldn't need a product owner, right, if the environment would be that it's

0:12:47.160,0:12:53.280
apparent that as a developer you can go directly to management and, well, share your ideas

0:12:54.240,0:12:58.680
and your your wishes, yeah. Okay, and I have one last

0:12:58.680,0:13:01.680
question before we go on to the next speaker and this is a personal question.

0:13:01.680,0:13:04.740
Okay. Was there any evidence in your study that

0:13:05.880,0:13:11.700
these things are easier to negotiate when management has come from a technical background

0:13:12.300,0:13:17.160
or are software engineers turned managers just as difficult to work with in this

0:13:17.160,0:13:23.340
sense as people coming in with with MBAs? Yeah, yeah, so based on my own experience

0:13:23.340,0:13:28.620
I would say that it works better indeed if the - if management has its - has a technical background

0:13:28.620,0:13:32.100
especially when it comes to the technical factors that I mentioned.

0:13:32.100,0:13:38.760
So I wanted to spend more time on code refactoring, code quality, software testing,

0:13:39.900,0:13:44.280
I think it's it's easier to understand the added value

0:13:44.280,0:13:48.900
of these factors if you do have a background in computer science.