MVP/MMF/MMR

Planning

Plan a marketable product, a feature and a release as soon as possible for quick feedback

Planning

Description

MVP/MMF/MMR are efficient ways to get some feedback loop for continuous improvement and learning.

The first thing to understand here comes from the MVP, the “Minimal Viable Product”. The MVP idea is to build a very basic product around a few hypotheses or features captured from the market and to deliver it ASAP to check those assumptions [Ries 2011]. From the collected results, some insight must then be inferred: either the hypotheses were valid and the MVP can be enriched or the whole delivered product is simply thrown away [SAFe 2021-34].

The concept of MMF (Minimum Marketable Feature) is merely the same thing except this is at feature level and improvements are usually, UX Designers are responsible for those MMF once approved [SAFe 2021-35].

This very same idea has also been applied to releasing (MMR) and yet another minimal thing with the MLP, the “Minimum Lovable Product” which is all about providing something that stands out and people love more than competitors [Merryweather 2020]. MLP can be much understood when comparing it to the Kano diagram [Moustier 2019-1]:

MVP vs MLP with Kano [McCahill 2014]

Even if the concept is easy to understand, the whole concept relies on two things that are not obvious to handle

  • Reaching a minimalistic set of features and sticking to it
  • Throwing things away when the market is not aligned with what was offered - this cognitive bias is actually named “loss aversion” [Wikipedia 2021] [Moustier 2019-1]

Impact on the testing maturity

To assess the assumptions made, the organization needs to measure how the market reacts to the proposed MVP. This can be done with polls or directly thanks to sensors coded inside the product. Those sensors then provide observables that tell about how delighted users are with the solution.

When it comes to MMF, sometimes the organization hesitates between several implementation options. In this case, A/B testing [Young 2014] [Moustier 2019-1] may be involved to let users decide which one is the best.

MMR should also be tested thanks to Canary Releasing. This technique enables users’ feedback on new releases without impacting the whole set of customers; in the worst case, only the few beta-testers would be disappointed.

Agilitest’s standpoint on this practice

If a MVP is targeted, the product should provide vital things. The drawback with those basic features is if they fail, discontentment is immediate; hence the quality standard should be high. An MVP should not be misconfused with a prototype, this is a product that is delivered to genuine customers. This infers that the minimal product should be correctly tested.

Moreover, the MVP will be the bootstrap to a bigger product, starting test automation from the beginning is a very good start because it avoids starting with some kind of technical debt that will probably never be refunded.

The good thing with Agilitest is the #nocode approach [Forsyth 2021] that enables fast scripting with few debugging, thus the testing automation cost is lower than automation with classical Selenium scripting and if the MVP should be thrown away, the loss aversion would be smaller.


Related cards

To go further


© Christophe Moustier - 2021