Here at Aampe, we’re working to change the way companies communicate with their customers. We do that through always-on experimentation. You create a bunch of messages: different attention-getters, different value propositions, different creatives, different products. Aampe will automatically put those variants into competition with one another, in varying contexts and schedules over time, to truly optimize your communication and generate valuable data about your customers that is entirely customized to your business.

There’s a lot to communication experiments other than the experiments themselves. We’ve built Aampe to let you make your own choices regarding these other parts. This post explains how.

I find it convenient to think about customer communication in terms of four buckets of capabilities:

  • Creatives/Copy. If you want to talk to a customer you have to actually say something. This includes writing an actual message, including a lot of message details such as font and subject lines, but can also include broader aspects of the creative such as images or videos.
  • Communication. If you want to communicate you have to actually send a message. This involves technical management of  email, sms, chat, push notifications, or any other channels.
  • Instrumentation. This includes any way you can measure a response to a message. This might include instruments customers don’t really notice (like link tracking), instruments customers notice but don’t recognize as instruments (like making a purchase or responding to a message), or could even include explicitly asking the customer for a response (such as a survey).
  • Automation. If you go to the trouble of creating a message, getting it to your customer, and seeing the response, you probably want to do something with that information. This category of capabilities includes any way you move data into or out of experiments, either from or into other business functions. This includes automated workflows, analytics, and things like that.

At Aampe, we’ve decided that the platforms that address primarily one or more of these four capabilities are potential partners, not competitors. We’re working to create integrations for every one of those products. We want Aampe to become a way for you to get even more value out of your presentation, communication, instrumentation, and automation platforms.

Some of the companies that fall into one or more of those buckets also offer experimentation. To be frank, in every case I’ve seen, experimentation is just a feature - one among many. And, to be even more frank, I haven’t been impressed with the experimentation feature in the platforms I’ve looked at. That’s part of the reason we founded Aampe. Experimentation, done right, can offer unparalleled value and growth. Experimentation, done wrong, can waste a lot of resources and cause a lot of frustration.

The only companies that might really serve as an alternative to Aampe are those who do experimentation more than they do any of the four other things. These seem to fall into two categories: web experimentation and other kinds of experimentation.

One category is web experimentation. This is a topic for an entire blog post in itself. There are a lot of companies that do website experiments. Many of them are impressive. But they all require a couple things: you need a web property that acts as your most important point of contact with your customer, and you need lots and lots of volume, because a huge sample size is the only way to draw trustworthy conclusions from unconditioned experiments.

I’ve noticed that many of these companies build a lot of their automation and instrumentation stack themselves rather than partnering with companies that do those things full time. These companies also, by design, offer very narrow capabilities for content creation and communication, because they’re really only interested in website content. So they end up putting their customers in the position where they have to choose specific automation and instrumentation tools if they want to also use the experimentation tools.

I put all other experimentation offerings into a generic “other” category. This might be communication experimentation, which is what Aampe offers. It might also be experiments with physical things, such as companies who arrange pop-up stores to test out business models, or who engineer prototypes to test product-market fit. I put all of these companies in the same category, because they all have one major thing in common: they’re all consultancies. Some of them do really impressive work. With Aampe, we’ve chosen to take a different direction.

We believe experimentation is now where machine learning was a few years ago when DataRobtot,, and Dataiku, not to mention the "auto-ml" offerings of Google, AWS, and Microsoft, became a thing. What all of these companies realized was that machine learning was well-enough understood to automate its use so, in many cases, people could get value from it without being experts in it. The same is true of experimentation: you don’t have to be an expert in experimental design to run a well-designed experiment. Aampe automates that expertise for you.

That’s what makes it possible for you to start using Aampe without any consultant. And without any consultancy fees.

We’ve built Aampe to serve as a sort of switchboard to connect all of the tools you’re used to using to communicate with your customers. You can take what you already like, and make it work better for you.