The Crux #6: Time travel only goes one way

State of the cloud news, tech company markets movements, plus a bunch of chip fab news this week.

This week’s tip is about asking customers about their life before they buy your product. Memory is fallible, and time travel only goes one way, so you only get one chance to do this.

Things to note

HashiCorp have released their second annual State of the Cloud Strategy Survey. I am puzzled by the headline stats where 90% of those surveyed say multi-cloud is working, but 94% are also wasting money in the cloud.

Cohesity has hired Sanjay Poonen, most recently of VMware, as CEO. This looks like another pre-IPO manoeuver, so I expect we’ll see an S-1 soonish.

Cloudflare reported earnings and Cloudflare predicted increased revenues next quarter so the line went up. The increase in loss doesn’t seem to bother markets for reasons I don’t understand, which is probably why I’m not a stock analyst.

Twilio came in under expectations which made the line sad. Twilio lost slightly less money (as % of revenue) than a year ago, unlike Cloudflare which lost more, so the lesson here is to lose lots of money, apparently.

Softbank’s vision Fund has lost two more execs, now more than 10 gone.

Japan is going to open an R&D centre for 2nm chip tech in partnership with the USA, by the end of the year. I assume Intel won’t be involved because then it won’t open until 2024 and be a 5nm process.

To understand more about Intel’s woes, this in-depth opinion piece at The Register is a good summary. Gelsinger has a lot of work to do.

It looks like the USA will ban selling advanced chip software to China. In unrelated news, industrial espionage has intensified for some reason.

Remember when law enforcement were doing lots of media about the An0m phone sting? Lawyers in Australia are challenging the lawfulness of how the messages were collected. One to watch, given the ongoing cold war on encryption.

Amazon is going to buy iRobot for $1.7bn, adding yet another surveillance device into company’s growing Skynet panopticon.

Someone built a 3D model of Ferris the Rust programming language mascot. Carcinisation will ultimately make all programming language mascots crabs.

Finally, someone has created a novelty language called Aussie++ that actually works. Americans, and others who find salty language challenging, might want to avoid clicking on this one.

Longer reads

A couple of longreads this week:

A really accessible explanation of how Kerberos works by Steve Syfus, a senior developer on the Windows Cryptography, Identity, and Authentication team at Microsoft (@SteveSyfuhs on Twitter).

A bonus longread, because it’s not that long, called The Bitter Lesson by Rich Sutton, distinguished research scientist at DeepMind and a professor of computing science at the University of Alberta. He argues that search and compute power wins over trying to replicate human ways of thinking, but I’m not sure I agree with his “the world is too complex to understand so we should give up” thesis.

Weekly Tip

Part of product management is understanding what customers actually care about, rather than what they say they care about. This can be hard, because customers lie; to you, and to themselves. We all do it, it’s just how humans work. Acknowledging this aspect of reality is important, because then you can use this knowledge to adjust what you do in response.

It’s particularly important when you’re developing your go-to-market. What customers think they’ll like about your product is often very different to what they actually like about it once they’ve used it for a while. When I do customer interviews, I like to ask questions designed to uncover this. It can be tricky because we humans have terrible memories and we forget what it was like to never have had the product we now take for granted.

To help with this, it’s often a good idea to start asking prospective customers about their life now, before they’ve bought your product, to establish a good baseline of what changed after they started using your product. Then, after the Proof of Concept (or, say, six months post-purchase) ask them what life is like now, and what they’ve noticed has changed.

Ask them what they remember thinking before they stared using your product. You, and they, may be surprised to discover how different their memory is from what they said at the time.

Then ask them to imagine life without your product now they’ve used it. You want them to be horrified at the thought. If they aren’t, your product isn’t really that valuable to them.

If you do this often enough, you’ll have some lovely statistics about the general outcome of using your stuff. You can talk about these outcomes with new prospects that are curious about what results they might get. You can explain why they’ll find your product invaluable, using concrete details from the stories that other customers have told you.

You’ll also be building a solid evidence base for what actually convinces customers to convert from prospects to buyers, rather than what people claim will make them buy when you give them surveys.

It’s really hard to go back in time to capture the real “before I was a customer” experience. Adding it to your onboarding experience can be a good idea, provided it doesn’t add too much friction.