Nowadays, in Software Development Contracts the key elements determining the price ... The main goal of the measures is
2°°International Conference on IT Data collection, Analysis and Benchmarking Tokyo (Japan) - October 22, 2014
Software Rates vs Price of Function Points
An Updated Cost Analysis
Rafael de la Fuente Founder and CEO of LEDAmc Insert here a picture Raúl Fernández SW Productivity Consultant
http://itconfidence2013.wordpress.com
Software Rates vs Price of Function Points: A cost analysis
Goals of the presentation
Nowadays, in Software Development Contracts the key elements determining the price are: the Rate and the Effort. The possibility to measure the quantity of software produced (the size in Function Points) allows us to assess whether there is a logical connection between the price of the projects and the software actually produced. The main goal of this presentation is to show what happens with the FP price in several scenarios:
G1. Implementing estimation models. G2. Implementing productivity models. G3. Introducing new competitors in a productivity model. G4. De-localizing providers.
IT Confidence 2014 – October 22, 2014
http://itconfidence2014.wordpress.com
2
Agenda • • • • •
Background Objective 2014 vs 2012 benchmarking Case studies Conclusions
IT Confidence 2014 – October 22, 2014
http://itconfidence2014.wordpress.com
3
Background Over the last seven years, LEDAmc has managed the productivity of more than 12,000 development projects of 10 significant clients in Spain (mostly telecommunication and financial companies). They were mainly small enhancement projects. The main goal of the measures is to control large contracts of Adaptive Maintenances, which involves the highest percentage of our clients’ development budget.
In 2012 LEDAmc developed a benchmarking study with some of these projects (we selected 3,405). The purpose of the study was to analyze the relationship between the rates of the different software providers and the price of the Function Point that they were offering their clients. This study was presented: •In
the UKSMA conference in October 2012
•An
updated study, in the IT Confidence Conference in Rio, in October 2013.
IT Confidence 2014 – October 22, 2014
http://itconfidence2014.wordpress.com
4
Background The basic numbers of the 2012 sample were: •10
Clients
•14
Providers
•3,405
Projects
•196,356
UFP
•2,168,192
Hours
•69,926,907
Million Euros
And the main conclusions were: Pressure There The
to lower rates ends up in lower productivity and higher FP price.
is not a logical relationship between Rates and FP Price.
performance of the providers changes dramatically among clients.
Clients
with only one development provider have the highest FP Price. IT Confidence 2014 – October 22, 2014
http://itconfidence2014.wordpress.com
5
Background 100
90
80
70
60
50
40
30
20
10
0 1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
CLIENT-PROVIDER Pressure to lower rates ends up in lower productivity and higher FP price. IT Confidence 2014 – October 22, 2014
http://itconfidence2014.wordpress.com
6
31
Background 100
90
80
70
60
50
40
30
20
10
0 1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
CLIENT-PROVIDER The same provider for different clients, offers significant differences in FP price. IT Confidence 2014 – October 22, 2014
http://itconfidence2014.wordpress.com
7
31
Background 100
90
ONE-PROVIDER CLIENTS
80
70
60
50
MULTI-PROVIDER CLIENTS
40
30
20
10
0 1
2
3
4
5
6
7
8
9
10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
CLIENT-PROVIDER The price of the FP with only one provider is higher. IT Confidence 2014 – October 22, 2014
http://itconfidence2014.wordpress.com
8
Objective LEDAmc has been continuously working for most of those clients, implementing Software Development Productivity Models over the last two years. Our productivity benchmarking data base has grown: It has more data for the same clients and new clients have been included. The main goal of this presentation is to show what has happened in these two years with clients who have worked hard to improve their software development productivity. We
will first explain the general evolution of the Function Point Price for those clients,
comparing 2012 and 2014 benchmarks. Afterwards, Finally,
we will focus on three case studies that are especially significant.
we will draw conclusions on the economics of software management, based on
our results.
IT Confidence 2014 – October 22, 2014
http://itconfidence2014.wordpress.com
9
2014 vs 2012 benchmarking The 2014 sample were: •Clients
that were in the 2012 study and have implemented some estimation or
productivity control •4
Clients
•12
Providers
•883
Projects
•44,897
UFP
•415,768
Man Hours
•15,428,772
Million Euros
What has happened with the FP price?
IT Confidence 2014 – October 22, 2014
http://itconfidence2014.wordpress.com
10
2014 vs 2012 benchmarking
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
CLIENT-PROVIDER The rates have remained almost the same. IT Confidence 2014 – October 22, 2014
http://itconfidence2014.wordpress.com
11
2014 vs 2012 benchmarking
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
CLIENT-PROVIDER The FP price is more homogeneous in 2014 than 2012. IT Confidence 2014 – October 22, 2014
http://itconfidence2014.wordpress.com
12
2014 vs 2012 benchmarking Rates vs FP Price in 2014
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
CLIENT-PROVIDER Apparently there is no relation between FP price and rate… IT Confidence 2014 – October 22, 2014
http://itconfidence2014.wordpress.com
13
2014 vs 2012 benchmarking Rates vs FP Price in 2014
1
2
3
4
5
CLIENT A
6
7
8
9
10
CLIENT B
11
12
13
14
15
CLIENT C
16
17
18
CLIENT D
.. but if we group the data by client, the results look different. IT Confidence 2014 – October 22, 2014
http://itconfidence2014.wordpress.com
14
2014 vs 2012 benchmarking Rates vs FP Price in 2014
P1
1
P2
2
P3
3
4
CLIENT A
P3
5
6
7
P2
8
9
10
CLIENT B
11
P1
12
P3
13
14
15
CLIENT C
16
17
18
CLIENT D
There are big differences in the FP price for the same provider with different clients. IT Confidence 2014 – October 22, 2014
http://itconfidence2014.wordpress.com
15
2014 vs 2012 benchmarking Conclusions The results are homogeneous on a client basis because they have implemented estimation or productivity control models that allow them control the FP price. Because there is no transparency in software development and the public FP price does not exist in the market, the FP price stills remains very different among the clients. The natural differences of the systems architectures and processes should not imply such FP price differences.
… now we will present some case studies highlighting the change that has been produced for these clients who have introduced control systems.
IT Confidence 2014 – October 22, 2014
http://itconfidence2014.wordpress.com
16
Case studies Case Study 1 Competition among providers + Estimation Model
INITIAL SITUATION OF THE CLIENT (2012)
High FP price
Only one provider
No estimation control
IT Confidence 2014 – October 22, 2014
http://itconfidence2014.wordpress.com
17
Case studies Case Study 1 Competition among providers + Estimation Model INPUTS (2013)
During 2013, the implementation of a centralized mechanism of estimation control has allowed to objectively contrast the costs offered by providers with corporate cost strategies.
It includes functional sizing, as well as other sizing methods (for parametrization, technical changes, test support,..)
Benchmarking studies helped establish a cost strategy and a target price for FP .
Introducing a new development provider fosters competitiveness among providers and helps establish a FP price within market values.
IT Confidence 2014 – October 22, 2014
http://itconfidence2014.wordpress.com
18
Case studies Case Study 1 Competition among providers + Estimation Model 1
INPUTS (2013)
2
1.
The rate was only reduced by 5%
2.
New development provider worked with a similar rate
TOTAL
1 1 PROV
IT Confidence 2014 – October 22, 2014
2
PROV 2
http://itconfidence2014.wordpress.com
19
Case studies Case Study 1 Competition among providers + Estimation Model 1
INPUTS (2013)
2
1.
The rate was only reduced by 5%
2.
New development provider worked with a similar rate
TOTAL
1 1 PROV
2
PROV 2
The FP price was reduced by 44 %
Similar FP Price among providers
TOTAL
1 1 PROV
IT Confidence 2014 – October 22, 2014
2
PROV 2
http://itconfidence2014.wordpress.com
20
Case studies Case Study 1 Competition among providers + Estimation Model
FINAL SITUATION (2014)
A good estimating process was used for all projects
44% reduction in FP price
More than 10 Million euros saved
FP Price within market values
Two competing providers in 2013 – Four in 2014
IT Confidence 2014 – October 22, 2014
http://itconfidence2014.wordpress.com
21
Case studies Case Study 2 De-localization of providers INITIAL SITUATION OF THE CLIENT (2012)
1
2
3
4
5
6
7
8
9
Specialized providers
On-site development
High rates and very low FP price
Estimation and productivity control implemented
10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
IT Confidence 2014 – October 22, 2014
http://itconfidence2014.wordpress.com
22
Case studies Case Study 2 De-localization of providers INPUTS (2013)
The client strategy was to proceed to the de-localization of the providers to a typical outsourced model introducing new development providers with lower rates.
The client also decided to create an in-house development team in order to preserve and extend the know-how.
IT Confidence 2014 – October 22, 2014
http://itconfidence2014.wordpress.com
23
Case studies Case Study 2 De-localization of providers INPUTS (2013) 1
2
1.
New provider with lower rate
2.
In-house development rate
TOTAL
1 1 PROV
2 2 PROV
IT Confidence 2014 – October 22, 2014
3PROV
3
4 PROV 4
http://itconfidence2014.wordpress.com
24
Case studies Case Study 2 De-localization of providers INPUTS (2013) 1
2
1.
New provider with lower rate
2.
In-house development rate
TOTAL
1 1 PROV
2 2 PROV
3 PROV
3
4 PROV 4
Logical behavior of FP price
In-house developments at lower FP cost than in
TOTAL
1
PROV 1
2
PROV 2
IT Confidence 2014 – October 22, 2014
3
PROV 3
4
PROV 4
outsourced scenario
http://itconfidence2014.wordpress.com
25
Case studies Case Study 2 De-localization of providers FINAL SITUATION (2014)
FP price controlled
Outsourcing process controlled
2014 situation
1
2
3
4
5
6
7
8
9
10 11 12 13 14 15 16 17
IT Confidence 2014 – October 22, 2014
18
19 20
21
22 23 24 25 26 27 28 29 30 31
32
33 34
http://itconfidence2014.wordpress.com
26
Case studies Case Study 3 Controlling large maintenance contracts
INITIAL SITUATION OF THE CLIENT (2012)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
IT Confidence 2014 – October 22, 2014
15
16
Ratio of the rates: 1.1
Ratio of the FP Price: 2.3
17
18
19
20
21
22
23
24
25
26
27
28
29
http://itconfidence2014.wordpress.com
30
31
27
Case studies Case Study 3 Controlling large maintenance contracts INPUTS (2013)
During 2012 the client implemented a productivity control for maintenance development.
An annual productivity SLA for all the providers was established.
The client also implemented a productivity and quality providers ranking.
Every two months the client presented to the providers their position in the ranking.
IT Confidence 2014 – October 22, 2014
http://itconfidence2014.wordpress.com
28
Case studies Case Study 3 Controlling large maintenance contracts INPUTS (2013)
Renegotiation of rates Ratio 2012: 1.1 Ratio 2014: 1.0
TOTAL
1
2
3
4
IT Confidence 2014 – October 22, 2014
5
6
http://itconfidence2014.wordpress.com
29
Case studies Case Study 3 Controlling large maintenance contracts INPUTS (2013)
Renegotiation of rates
Ratio 2012: 1.1 Ratio 2014: 1.0
TOTAL
1
2
3
4
5
6
FP cost control Ratio 2012: 2.3 Ratio 2014: 1.7
TOTAL
1
2
3
4
IT Confidence 2014 – October 22, 2014
5
6
http://itconfidence2014.wordpress.com
30
Case studies Case Study 3 Controlling large maintenance contracts FINAL SITUATION (2014)
Uniformity of FP price
9.5 million euros saved
2014 situation
1
2
3
4
5
6
7
8
9
10
11 12
13
14 15 16 17 18 19 20 21 22 23 24 25 26
IT Confidence 2014 – October 22, 2014
27
28 29
30
31
32 33 34
http://itconfidence2014.wordpress.com
31
Conclusions 1. It is relatively easy to influence FP price. The main levers are: >>> A centralized estimation control >>> A development productivity control >>> Several providers in competition
2. Working on those key elements the results are much better than working only on the rates: >>> The behavior of the unit cost of production (FP price) is more homogeneous. >>> The cost of the development process can be controlled.
3. Diversification of providers with joint management results in a better uniformity of the FP price. IT Confidence 2014 – October 22, 2014
http://itconfidence2014.wordpress.com
32
![Software Rates vs Price of Function Points: A Cost ... - WordPress.com [PDF]](https://m.moam.info/img/260x300/software-rates-vs-price-of-function-points-a-cost-_647cee78098a9ebc738b4573.jpg)
![Rates vs Cost per Function Point - WordPress.com [PDF]](https://m.moam.info/img/260x300/rates-vs-cost-per-function-point-wordpresscom-pdf_6480ec58098a9ea8138b45e3.jpg)
