Project Management

Please login or join to subscribe to this thread

Velocity tracking when managing various technologies

linkedin twitter facebook   Estimating  
avatar
Hernan Voto Project manager| Rockstar Coders Caba, Buenos Aires, Argentina
Hi,
The scenario is this: A project backlog is composed mainly by 2 types of cards (sometimes 3) Frontend, Backend and Misc (eg UX)
To add to my problem, I have to support 2 different platforms (Android and iOS) plus a RoR backend.

Some cards are for iOS, others for Android and sometimes are all for one platform only.

For instance I may have during one sprint 2 iOS cards 3 backend cards, 1 UX, and then for the next 2 sprints 75% more Android cards than iOS.

As a result, the velocity is useless in my project. Maybe this type of project cannot benefit from it.

How would you track it properly?

thanks
Sort By:
avatar
Wade Harshman Scrum Master| GDIT Indianapolis, In, United States
Hernan,

There is no requirement to track velocity, and it should only be used for your team, anyway. If tracking velocity does not help your team, then the practice is waste and should be eliminated.

However, you can continue to track overall team velocity, even though there are different types of work to do. If your team's iteration planning is velocity based, you should consider average velocity because it isn't as volatile. In some iterations, you can accomplish more. In other sprints, you will accomplish less. Over time, though, the average velocity should not change a great deal, so you can still use it to forecast.

If certain types of work are more difficult for your team (if you don't have front end developers, for example), then the team should increase the estimates for those items. If it's more a matter of capacity- for example, if you have 3 back end developers but only 1 front end developer- then track your average velocity for long-term forecasting but switch to a commitment driven approach when you're in iteration planning. This takes longer than velocity driven planning, but it produces better results since velocity changes from one iteration to the next.
avatar
Sergio Luis Conte Helping to create solutions for everyone| Worldwide based Organizations Buenos Aires, Argentina
If you are talking about backlog composed by user stories the first error is to include user interface into user stories. User stories must not be User Interface related.
avatar
Hernan Voto Project manager| Rockstar Coders Caba, Buenos Aires, Argentina
Thanks Wade, makes sense.
avatar
Hernan Voto Project manager| Rockstar Coders Caba, Buenos Aires, Argentina
Sergio, I know that although to be sure of what you mean, lets understand your comment: stories in our backlog (and 2 previous companies I worked for) are defined for frontend and backend separately, since despite theory, in practice for mobile projects there are no full stack dev available (at least in my company)

Are you suggesting to have more generic stories, use the combined estimate and then use tasks to separate/assign backend frontend tasks?
...
1 reply by Sergio Luis Conte
Mar 21, 2019 11:05 AM
Sergio Luis Conte
...
No me dí cuenta que eras compatriota!. Solo como comentario siempre contesto mas allá de la teoría. Tuve la oportunidad de trabajar con Kent Beck y Allistar Cockburn asi que te comento lo que hacíamos porque a lo mejor te sirve. Las user stories son una invitación a seguir hablando de lo que pretende el usuario sobre el producto o como prefiero decir del sistema (sistema puede no ser software) cuando interactúa. Ahora bien, de allí se derivan las distintas capas del sistema pensando en arquitectura. Lo que haciamos nosotros era, en caso que una de las capas fuera la interface, es bajar a detalle usando Use Cases. Con esos Use Cases construíamos. Es decir, una cosa es lo que el cliente o stakeholder (como Uds. lo llamen) pretende y otra es construir pensando en la arquitectura. Cuando pensan en la arquitectura es cuando hay que separar las capas, ya sea uses Use Cases o Stories para seguir detallando qué se espera de la solución. El tema velocity para lo único que sirve es para ajustar la distribucion de User Stories entre los sprints. Pero el problema es que ya si utilizas Story Points como si medis Velocity el error inherente que tenes en ambos casos es muy muy grande. Recién vas a poder "confiar" luego de no menos de tres iteraciones utilizando esos métodos de estimación. No sé si con eso te ayudo.
avatar
Kiron Bondale Retired | Mentor| Retired Welland, Ontario, Canada
Hernan -

Rather than break your stories out by layer of the stack, why not take a slice of end-to-end functionality and have the team deliver that (back end through front end)? That way, you'll have multiple stories each sprint which represent "real" value, and velocity starts to have meaning (e.g. how many small valuable stories can we complete?)

Kiron
avatar
Hernan Voto Project manager| Rockstar Coders Caba, Buenos Aires, Argentina
Kiron, thanks!

is that different from what I answered Sergio? just to be sure...
avatar
Sergio Luis Conte Helping to create solutions for everyone| Worldwide based Organizations Buenos Aires, Argentina
Mar 20, 2019 4:28 PM
Replying to Hernan Voto
...
Sergio, I know that although to be sure of what you mean, lets understand your comment: stories in our backlog (and 2 previous companies I worked for) are defined for frontend and backend separately, since despite theory, in practice for mobile projects there are no full stack dev available (at least in my company)

Are you suggesting to have more generic stories, use the combined estimate and then use tasks to separate/assign backend frontend tasks?
No me dí cuenta que eras compatriota!. Solo como comentario siempre contesto mas allá de la teoría. Tuve la oportunidad de trabajar con Kent Beck y Allistar Cockburn asi que te comento lo que hacíamos porque a lo mejor te sirve. Las user stories son una invitación a seguir hablando de lo que pretende el usuario sobre el producto o como prefiero decir del sistema (sistema puede no ser software) cuando interactúa. Ahora bien, de allí se derivan las distintas capas del sistema pensando en arquitectura. Lo que haciamos nosotros era, en caso que una de las capas fuera la interface, es bajar a detalle usando Use Cases. Con esos Use Cases construíamos. Es decir, una cosa es lo que el cliente o stakeholder (como Uds. lo llamen) pretende y otra es construir pensando en la arquitectura. Cuando pensan en la arquitectura es cuando hay que separar las capas, ya sea uses Use Cases o Stories para seguir detallando qué se espera de la solución. El tema velocity para lo único que sirve es para ajustar la distribucion de User Stories entre los sprints. Pero el problema es que ya si utilizas Story Points como si medis Velocity el error inherente que tenes en ambos casos es muy muy grande. Recién vas a poder "confiar" luego de no menos de tres iteraciones utilizando esos métodos de estimación. No sé si con eso te ayudo.

Please login or join to reply

Content ID:
ADVERTISEMENTS

The only people who find what they are looking for in life are the fault finders.

- Foster's Law

ADVERTISEMENT

Sponsors