Lean Startup in big corporations - Lean Startup Machine Talk
The main problems in big companies are well known of product managers and entrepreneurs.
loss of ownership and loss of empowerment (it's like nobody is never responsible for anything),
which all leads to big cycle and lead time.
We had all those problems when we started to build chooseyourboss.com within the biggest french web based classifieds company. our marketing team was working for 4, or 5 products at the same time. Their role wasn't always clear : were they working on the product? On the growth only?
Our sales were in the same situation : they had a big product making millions, and for most of them, were afraid to sell a new product without any 'proof' that it was going to make it.
I imagine that a lot of product managers and intrapreneurs have seen the same situations.
What helped us is a framework we crafted during 18 months, that we called experiments. Experiments are testable and measurable actions that allow to validate or invalidate assumptions and hypothesis we make everyday as product managers.
Who are we ? built 2 startups, failed one, one side, one building.Trying to use Lean Startup practices since 3 years, after reading Running Lean.Right now : building chooseyourboss as intrapreneurs, building tools for product managers.Important : we are not ayatollahs. At beginning, we were like “monkey see, monkey do”, now we’re a little more souple =) What is important to understand is that every context is different, every market, every product is different, so what we managed to do doesn’t necessarily apply to you. But we believe that there are some lessons here. So don’t forget : there is no “absolute truth”. Lean Startup just like Agile or Lean is a way of thinking, a discipline, not predefined rules to make you successful or not if you don’t follow them.
What we strongly believe in, after 3 years of practice in our own ventures, and also as product manager in big companies.
From the idea to the dollars, or at least the lessonslearned
As Lean intrapreneurs/Product Managers what do we want?Reduce the market risks
It is important to note that we started, at the really beginning, to build a minimum viable team, with all the skills needed to implement the product, from business to engineering
We needed to align this team around common measurable goals, and make them follow those ones, even after the “ticket” is “done”.We needed to kill this “DONE” paradigm. And give birth to a new “done
What we gained here?The measurable aspect of those experiments gave ownership to people working on the project. They understood that their actions will be checked, and a ticket done was ok, but done needed to be “KPI compliant”.Problem, no following, no improvements, but high power inshareability and understanding, + + lessons learned. A lot of informations were needed. Here for example, the marketing was looking at us saying : what was done exactly? It was more, by us, for us. NOT GOOD YET =)
This was the slide we were sharing to everyone every Thursday, available on dropbox. This was the high level view, with : The problems we were solvingThe segmentThe type of experimentation (quali/quanti)
What we gained here?Here, there is an indeniable testable aspect. Expliciting clearly the problem, having to write in a small box the big picture solution, and also, the mesurable goals made the people think small, testable and try things differently. This aspect helped us focus and align the team around credible and written goalsThe main problem here, is that it was definitely lacking follow up, and be able to understand what action change what metric sometimes. Plus, this was still too clunky.In ticketing organisations, one day you talk to someone, one other day to another one, for a various number of reasons. And while some experimentations were lasting weeks/months for some of them, people were coming back and saying we should do something we already did weeks back, and we were on for explanations why we wouldn’t do that, but without the “data proofs”, and the lessons learned of that time, which would have been seriously helpful.
What we gained here?The fact that you need to follow those “experiments” through time, weeks, sometimes months helped us empower people and make them accountable and responsible for the whole product. They understood that we were iterating on everything, constantly trying to improve, or kill, or create new things that would bring value to the customer. They were definitely still in the ticketing culture, but for operational purposes (who does what), but not for product management anymore.This definitely helped us to implicate the marketing way more than the other solutions, because anybody could come on the experiment and understand what happened, through time, and quite easily. Every action was linked to an incomeThis is not only actions on the mail but on the product itselfWe follow this mail + the landing + the high metrics > everything was shared with the sales to build up powerful presales speech for instance
What we achieved?
Butdon’t forget =)
Lean Startup in big corporations - Lean Startup Machine Talk
Lean Startup and innovation inbig corporationsHow experimentations helped us align dispersedteams, and assert our assumptions fast.
We built a team with the minimum set of skillsto tackle the problems.
No follow up. No measurement. No lessons.No improvements.
[ex·per·i·ment] (/ikˈsperəmənt/) - noun"A scientific procedure undertaken to make a discovery, test a hypothesis, ordemonstrate a known fact."Synonyms : trial - test - try - attempt - essay - assay - tentative
Everything started with what we calledlessons learned. It was simple slide deckscontaining shareable information, as metricreports and objectives.
HypothèseLes candidats qui reviennent vont jusqu’au boutMesureConclusion1/3 de la base est réutilisable.=> Effet déceptif + Profil déjà périmé ?Mail On siteOpen Rate30 % 55%Mails240 240Conversion rateTunneld’inscription fini50% 40%Visits35 79Bounce30% 25%Click rate40 % 51%IndicateurPrévu Mesuré28/03
We needed something better, as shareable,but with more information, like the context,the big picture solution, the actions taken,the metric following, and still, the lessonslearned. We called this one“experimentation”.
Expérimentations en coursContexte (Problème, Risque, Idée)Acquisition SEM / SEO CANDIDAT RECRUTEURquali quantiOpportunitésDMR non faite par recruteur10% des candidats avec des offres Envoyer des opportunités aux candidatsCANDIDAT RECRUTEURquali quantiNouvelles offres commercialesOffre startup (com’ & positionnement produit)Offre essaiOffre de stage Offre cabRECRUTEURquali quantiCommerciaux offlinePas de réussites partagéesproblème de confiance dans la solutionmanque d’info au niveau des commerciaux, manque de feedback venant des commerciauxRECRUTEURquali quantiCommerciaux online1 seul canal d’acquisition commercial, Commerciaux OfflineManque d’offres, problème de confiance dans la solution Développement du canal OfflineRECRUTEURquali quantiConférence + MatchmakingAcquisition candidat + image de marqueMatchmaking / Conférence dans les écoles, les salonsCANDIDAT RECRUTEURquali quantiPage 2Page 3Page 4Page 5Page 6Page 7Plan de contactBase inactive / qualité des profilsRECRUTEURquali quantiPage 8
ExperimentationResponsable : VincentContributor(s) : Benoît Informed(s):Context (Problem, Risk, Idea)Envoyer offre en mode pro-actifPb : DMR non faite par recruteur, 10% des candidats avec des offresLessons learnedLes annonces Web Tech marchent très bienToulouse / Paris en têteLes opportunités marchent mieux que les DMR.Les candidats ne sont pas recontactésLes recruteurs n’arrivent pas forcément à contacter les candidats.Solution Big PictureEnvoie des matchings à 400 candidatsSi ils sont intéressés, on les propose aux recruteursGoalsAu 15/01Candidat avec une offresuggérée: 30%Candidats intéressés : 400Maiwenn : Tx d’ouvertureimprove_profile 40%Tx Rebond après improveprofile : max 20%Current Metrics434 dmr acceptées / 433intéressées (7/02)2079 acceptées dont 642 dmracceptées / 1437 intéressées(18/03)Maiwenn Tx d’ouvertureimprove_profile 49%; (11/04)+5%Tx d’ouverture doublé (11/04)Actions• Suggestion automatique• Plan de relance pour les opportunités automatisé• Suggestion sur le profil public• Améliorer le mail d’amélioration de profil• Rediriger vers edition profil• Template + objet• Suggérer au candidat qui n’ont pas encore eu d’offres
The last thing we tried, that definitely is acomplement of the previous version, takesinto account the lack of follow up. It’s whatwe called experiments. And it worked for us,and definitely changed the way we workedas a team.
We decided to organize our team around“lessons learned” and “experiments”. And webuilt the tools to support this new organization.
The measurable aspect of those experimentsgave ownership to people working on theproject.The testable aspect helped us focus and alignthe team around credible and written goals.Follow experiments through time, weeks,sometimes months helped us empower peopleand make them accountable and responsiblefor the whole product.
This was a long process 9 months.We iterated. As everything in LS, weexplored, we pivoted to find what was rightfor us.Annex question : is the journey or the result what made our teamstronger?( you have 4 hours ;) )