The document outlines 5 tactics for experimenting in a corporate environment:
1. Run many small experiments in parallel teams to test ideas quickly.
2. Focus each experiment on testing a single hypothesis about customer problems or solutions.
3. Begin experiments by engaging customers to understand their needs and test potential solutions.
4. Gather enough data through customer interviews or other tests to determine if hypotheses are supported statistically.
5. Leverage all available corporate resources like marketing, IT, purchasing to conduct experiments at low cost and with real customers.
Yulia Smotrova - Running experiments like a pro in a corporate environment
1. Founder of “Time To Leap”
Official Lean Startup Ambassador - Berlin
Former Corporate Innovation Manager at PMI
11 years of working with innovative initiatives
MBA, Stanford Innovation & Entrepreneurship
Yulia Smotrova
19. 1. Go for N of experiments
2. One experiment per hypothesis
3. Start with a customer
4. Scale to statistical significance
5. Use all resources
20. Thank you!
+49 176 85610345
Time To Leap
yulia@t2leap.com
http://t2leap.com
Let’s
Build -
Measure -
Learn
together
Editor's Notes
Hi my name is Yulia and i was a corporate innovation manager for 4 years at a multinational organisation. Building FastForward innovation program from scratch for 25 countries made me a true believer that Experimentation is the fuel of innovation, without meaningful information you cannot make decisions, and you only get this by building and measuring.
But running experiments in a corporate world is not a walk in park. Who has ever had problem raise your hands: confusing KPIs, delayed decision making, lack of resources, legal limitations. Was it painfil? I bet! Today I would like to share with you 5 tactics that helped me at my job and may help you!
First step. Throw away that ideation box and a one pitch session to decide on the future of the idea. These ideas will be full of assumptions. Go for number of experiments that test these assumptions and have at least 4 pitch sessions with decision makers to track the progress.
Set innovation team KPI as 10 experiments min per 2 weeks. Let innovators run experiments in parallel of two or three people per small team and combine the results at the end of the day to plan for the next steps together.
By incentivizing number of experiments over number of ideas, you fail fast and cheap and reduce the fear of failure by quickly killing even your dearest darlings. My teams cycled through 5 ideas per week and did not even feel bad about it. They enjoyed the feeling of progress coming from the experimentation.
The second tactic is not making Pilot your MVP. For 1 idea you can have 10 mvps testing 10 different hypothesis. If any one comes to you saying just built the whole thing and ship it to a few clients, use this technique. Ask this person. Imagine it did not work? Would you know why?
If the answer starts, perhaps…red alert. If you don’t know what killed your pilot or made sales grow, you won’t be able to make a decision on the next step to take your idea forward. Break down experiments per hypothesis, so you know the correlation, what strings to pull or what to cut.
Third tactic. No matter how lucrative it is to start ideating about the product and the budget to build it. Start with understanding a who will buy the product. Selecting your early customers first sets the rhythm of your experiments. In case of b2b you will need more time to get ahold of them and use an intent letter to measure an interest.
If your customer is internal, engage with external party to conduct some tests to avoid bias and you no longer can use money to check their interest, so you will use the effort as measurement of success. if a colleague is ready to spend 30 min to apply to use your solution, and after refers it, it is clear proof that the product solves a real problem.
And always ask a question, how can we test assumptions for free? Gain the speed of experimentation starting with 0 budget. Plan your budget only for the next experiment. Because if it fails, you only lose this amount of money. And if you prove your hypothesis, you already have real data to use to get more funding at the next pitch progress session.
Forth tactic - scale slow toward statistical significance. Normally first hypothesis should be tested by qualitative research just answering questions such as Who are our customers and what problem can we solve for them? And only then Can our idea solve it, can we build it and how can we make benefit from it?
It is ok to start with 2-5 interviews, just to get a feeling of a direction, later scale it to 30 interviews to cover 1 hypothesis. Statistical significance will take place later, when you start running additional experiments to strengthen your hypothesis. Don’t worry about big numbers too early.
In my personal experience we made sure to interview 10 people at the first day of the teams’ innovation journey. That approach gave us a good reality check and memorable aha moments. 70% of teams had to change their assumptions after talking to potential customers that early on.
Tactic number 5 - Use all the corporate resources. You are not alone and you have some good advantages working for a large organization. However beware if you start running the experiments very enthusiastically, don’t be surprised that the first knock at your door will be from legal.
Be proactive. Go to Legal and set the permission boundaries for experimenting in advance. Get clearance on what can be done with up to 10, 100, 1000, … people. Most likely you might need no approval to just go to a cafe and talk to people if you don’t mention your company or product.
After Legal the next stop should be Marketing. Don’t reinvent the bicycle doing a customer research, save your time. Marketing can help you with getting your questionaries right, provide data on market size and advise on the tools to run your market research.
The third stop is your IT department. Especially if your product is digital, you can already get advise on the limits of your IT ecosystem. They can help with licenses to cool programs and refer you to vendors and some freelancers to build MVPs fast.
The last and not least it is Purchasing. They were a lot of help for us. They can speed up the contract process, give you access to the list of vendors, you can contact for your b2b research. You might also get into agreement with a vendor and run combined experiments without compromising a brand.
Going for the Number of Experiments, testing 1 hypothesis at a time, always starting with your target customer and moving from qualitative to quantitative research, with using all your corporate resources can help you to ace lean startup way at a corporate environment and experiment like a pro.
Being an ambassador I couldn’t resist to put a bit of Eric Ries here to tell you one more time Build Measure Learn and you will find your way to continues innovation no matter of the size of the company.