Interconnected

Raw thoughts on Amazon Dash

Amazon Dash... yup, makes sense. Give away light-weight Internet of Things gadgets to encourage purchase of fast-moving consumer goods. Tiny plastic buttons that allow for instant product ordering: Your entire house is now a shopping cart.

We worked on a bunch of similar stuff at Berg -- such as Cloudwash where purchasing washing powder was part of the machine itself, and some secret projects where the transactions and connectivity explored different configurations. It was my colleagues who spent most time figuring out the service design and all the mini design interactions -- but I was steeped in this for a year or two, so I figured I would dash out some notes...

Dash. Ho ho.

Look, a warning. I haven't proofread these notes so they might make no sense. And there's nothing about how the Dash is designed, or what this means for your strategy. Feel free to get in touch for a coffee if you want to talk about either of those aspects, or if anything in what follows rings a bell for you...

It makes economic sense

By putting the Dash Button in the home, it lowers friction to purchase and shifts the distribution channel to Amazon.

This is good for:

  • customer retention through reduced churn. You're a loyal Tide customer, but you've run out. Instead of running to the nearest store and grabbing whatever brand is on sale today, you hit the button
  • customer retention through reducing the replenishment time
  • switching to a low-cost distribution channel. Cost of sale is likely lower with Amazon or direct sales, versus physical retail that needs to cover in-store inventory and staff

You could probably make a pretty simple equation out of this, saying something like... it's worth it when:

revenue from selling N buttons + N1 * channel cost saving + N2 * margin on additional sales + value of effective marketing > development cost + server cost for 1 year + cost of producing N buttons

where N1 is the number of buttons used by existing customers who switch channels, and N2 is the number of buttons used to make new purchases.

Assuming you're giving these buttons away, and assuming you don't have access to the marketing budget, the relevant costs become:

  • how much do these buttons cost to develop (not much if you're Amazon, because you already have all the bits from your other products)
  • how much do these buttons cost to run (not much if you're Amazon, because you already have the back-end servers and the logistics)
  • how much do these buttons cost to produce -- the component cost has dropped sharply over the last year or two. Note that the Dash Buttons are small and use battery powered wi-fi... they fall out of the smartphone supply chain. Or to put it another way, these buttons are the peace dividend of the smartphone war

And when these costs fall enough, the question becomes: Do you reckon you can make a profit on the 2% of customers who will buy 5% more a year?

It's the new marketing

I brushed over the marketing benefits above, but let's not forget that we're enamoured with web-connected physical things. Internet of Things gadgets are novelty; brands that launched with Amazon will be pleased with the publicity.

Though... why not go with a Youtube video for the publicity, and not bother making the thing itself, like Evian's fridge magnet launching their water subscription service? It's a lot cheaper.

I can think of a couple of marketing benefits to actually making and distributing the button itself.

Advertising is changing. You need Adwords and Facebook ads to fish where the fish are... but come on, they're boring, they don't create the emotional impact of a full-page glossy magazine ad, or a great TV spot.

So, for me, these buttons are the first step to a new channel. Sure, they're utilitarian -- they order more consumables. But they also tell a story that this brand is there for me, it's in my home, it's at hand, on-demand. Products are people too and this button has a character... it's now up to us to create physical objects that have different characters, one that might be more suitable for high fashion, fragrance, whatever.

You need these different marketing channels because building awareness requires a drip-drip approach... a banner ad here, a Buzzfeed list there, a button on the fridge. A new advertising component.

E-commerce continues to grow at the expense of physical retail, and discovery in e-commerce continues to suck. Another way to think about the Dash Button is that it's doing the same job as a shelf-end promotion, or a BOGOF. This is a replacement for the lost world of point-of-sale discovery tricks, and it's in-home.

In a way, we're really seeing the future of marketing here. We've separated awareness (advertising) and distribution (stores) for so long, but it's no longer the way. When you get a Buy Now button in a tweet you're seeing ads and distribution merging, and the Button is the physical instantiation of this same trend.

In the future that Simon Wardley paints, this is a warning shot against the supermarkets, and in the future every product will carry a buy button. We're already in this world with smartphone apps: Because App Store discovery sucks, the best mode of distribution is word of mouth. The more downloads you already have, the more downloads you'll get. Which is why app publishers need to get their dirty mitts on your address book. Is this a future we'll see with consumable home products, too?

Why Amazon?

The costs of making the button dropped for Amazon before anyone else -- as I said above, they have the technology Lego bricks already, and the logistics available. But there are two other groups I've run into repeatedly who have looked seriously at the same concept

Traditional manufacturers of either FMCG or the kit that consumes it (that is, makers of washing powder and makers of washing machines). In 90% of cases, these manufacturers don't know how to speak with end consumers. They maybe know how to speak with consumers pre-sale... but they tend to outsource that to marketing agencies. But typically their customers are distributors and retailers, not end consumers.

This is a problem because the Dash Button isn't just a button, it's a communications channel. It's an app. It's push notifications. It's the update emails. Somebody needs to write that copy, somebody needs to manage the feedback, somebody needs to choose when to do a price promotion.

For most manufacturers, there's not a group in the company that knows how to do this, so making the Button isn't an engineering challenge -- it's a difficult corporate re-org where the voice that would advocate that strategy isn't even present in the leadership team.

There's a particular problem for the device manufacturers, as opposed to the manufacturers of the consumables. Connectivity of shared objects means back-end servers, and these cost money to run. When you're a big company, you think via the profit and loss. A connected machine which just offers neat user features (e.g. a notification when your wash cycle has finished) doesn't make sense on the P&L: it just means you have to knock points off your margin when you first sell it, to put cash aside to keep the service running. The only way connectivity makes sense is to align the recurring costs with some kind of recurring revenue.

In 2015 that means creating a sales channel for consumables... which the white goods manufacturers don't have available to them. Ink-jet printer manufacturers do. And we won't be stuck as this stage forever. Give it 12 months, and you'll have a microwave with in-app purchase. I'm only half kidding.

The marketing agencies get it. But marketing agencies are organised (have teams; do sales; are paid) around campaigns. Campaigns have a beginning and an end. In the web world, we're used to the KPIs of ongoing services (acquisition, activation, retention, etc): we measure Monthly Active Users. That's what you need to build a new sales channel. But it's not how you run a successful campaign.

And besides, marketing agencies need to do something different every quarter -- they don't build up a tight ops machine of technology and logistics. Building the Dash Button isn't, for them, just the next easy thing to do.

On my list of startups I would do - which is way too long and honestly, mostly awful - is an agency that offers to speak with customers post-sale.

So the manufacturer of Break Maker X would get us in to run the communications channel with the customers... either a button on the machine or just the app. And we'd all come from the web world, so it would be all A/B testing emails, looking at the button stats, partnerships with legendary bakers to sell artisan yeast, all that nonsense. We'd run it on a retainer basis, and eventually get bought by one of the big networks for a billion quid.

By the way, there is one group I've met who do (a) know how to have a customer conversation, and (b) have an incentive to cross-sell and up-sell on that channel. And that's the retailers, especially the retailers who already have a logistics network and run post-sales activities such as an extended warranty programme.

In the UK, that means John Lewis/Waitrose -- I was half expecting to see this button come from them, first. But they don't have "retail platform" in their DNA like Amazon.

And there are some manufacturers who understand that the product is the channel. But I can't say who because I'm still under NDA.

Last thoughts

I have to say, the Amazon Dash Button makes a ton of sense to me, as the near future of the Internet of Things, in a way the internet fridge never has.

I've always said that if I was making an internet fridge, I'd just make a fridge magnet that ordered milk. Or, better, a tiny talking Fridgeezoo pet that - when tapped - would text both me and my wife "we're out of milk!"

It's that shared use that makes this button really great. Smartphones are still stuck in the Personal Computer era... but my shopping list isn't personal. We live with flatmates, families, and friends. We can hack it -- at home we use shared Reminders on iPhone for the weekly shopping. But my online grocery order is associated with a single user account. It seems dumb; the button fixes that.

It's the ability for some kid in the household to say "hey, need more toothpaste" that leads to the button requiring wi-fi: Bluetooth is cheaper, but it would need a paired smartphone around.

And I am curious to see how it all works. The service design is complicated... without a screen, or any feedback, maybe your order is already on the way because somebody else hit the button, but how do you know? Sure the button automagically doesn't place a repeat order until the delivery is made, but what about more complex orders, or what if what you want is out of stock? Ultimately, this'll do for 80% of the cases, which is more than enough.

And then - for the rest - there's the Dash Replacement Service, Amazon's button-to-logistics technology all wrapped up, and ready to be integrated into whatever product you're creating. I see Quirky are building this replenishment service into the device itself:

a new line of smart appliances including an artisanal pour-over coffee machine, a baby formula maker, and a pet food dispenser. Each appliance will measure remaining consumable supplies and place an order using DRS before running out.

DRS is the best bit. Can you imagine what the life-time value is, for Amazon, if [white goods manufacturer] bake this into a washing machine, an item with an 11 year replacement cycle?

Where next

We're seeing the various stacks line up to own the smart home ecosystem:

  • Apple... offer Homekit to chip makers and device manufacturers, make every product in the home a smartphone peripheral. We'll make it easier to get to market, but only if you commit to us.
  • Google... with Nest establishing a beach-head, go for a data and orchestration play, providing feature differentiation. I wouldn't be surprised to see a Homekit-like play.
  • Samsung... leading the charge for the white goods manufacturers with SmartThings and the open ecosystem it promotes.
  • Amazon... end-to-end technology, combining both hardware connectivity through to back-end servers, payments, and logistics. Give it a year, then make the underlying chips available to any developer.

I know the Dash Button feels utilitarian -- but to my mind it's also eminently sensible, and executed in exactly the right way. I like it. If it does well, it'll make possible connected devices which are less utilitarian. I'll like that even more.