SF Design Week is Awesome!
A while ago, I attended a few more events that were really interesting. The first was a Growth Hacking talk from Recurly, which actually wasn’t a part of the SFDesignWeek events but still really interesting, and the second was a talk about Data Visualization from Salesforce.
Driving Company Growth Through Testing
I learned a lot at this talk, even though I probably wasn’t the target audience. I think it was probably more for the management type, but it was still really relevant to my interests. It was all about best practices when using testing to drive growth. Sean Ellis, current CEO of GrowthHackers.com and marketing leader of breakout companies including Dropbox and LogMeIn (IPO), was the main speaker for the talk. He did a presentation, and then it was followed by an interview from a Recurly exec, and then Q&A with the audience.
The Process Cycle
Prior to developing a process, you need to identify your “must need” users — users that think your product is a must-have for the long-term. These are users that answer “very disappointed” when you ask them “How disappointed would you be if you could no longer use this product?” These are users that would convert to a premium model. Sean described a pretty clean process cycle for businesses to follow when testing:
- Unbridled Ideation
- This is the stage where you spit out as many ideas as possible, viable or not. Sean suggested not relying on one person’s ideas, ideating across all areas (like acquisition vs. revenue), and thanking contributors personally and providing incentive.
- Work through a constructive critique, and be transparent about why an idea may or may not work (takes too much time, etc)
- Prioritize Backlog
- This is where you formalize the ideas created in step 1 in an “Experiment Doc,” and include preliminary research, a hypothesis (every test should have one), and a target lever.
- The goal for every test is the highest impact with the least effort.
- High Tempo Testing
- Sean asserted that “tempo is everything” and suggests having weekly meetings and a test launch target. These weekly meetings should never include conversations that include “wouldn’t it cool if…?” because that just takes forever and nothing gets done. I loved this, actionable items are the goal at this stage.
- Constantly assess the process and try and fix bottlenecks that are preventing test tempo.
- Build Knowledge Base
- Every test leads to learning, and this needs to be documented. This ensures transparent access and sharing.
There were a few things that Sean mentioned that stood out as pretty interesting to me.
- Many changes across pricing led to the conclusion that people aren’t very price sensitive. Decreasing price of his service didn’t increase demand, but in fact increasing it did.
- Sidekick at Hubspot runs 30 tests a week, which is crazy. I use and love this product as well.
- Moving the email signup form from the bottom of the page to the top caused a 700% increase in signups
- You can’t always be clear on cause and effect when running tests, you can measure like “I’m 80% sure than we had results because of the test we ran.” You also use predictive actions like “someone who comes back 10 times in the first week is more likely to come back in 10 months.”
- You can using a negative hypothesis at times, like turning off Adwords for a week.
- Embedded media (ex: modal video) adds more engagement time. This was relevant to a project I’m working on at my internship.
- Testing teams work best with 1 growth engineer, 1 growth designer, and 1 growth project manager.
This was a super interesting event and I learned a lot that will be really useful in the future! Stay tuned for a recap of the Data Viz event at Saleforce.