How to Create Last Mile Campaigns

Learn to sell more in the last mile

Last mile marketing is something of a science. Campaigns begin as ideas, which are tested, and then refined. This requires a diligent, methodical approach. To create campaigns, you first need to hypothesise and prioritise.


Every Amigo campaign is hypothesis-driven.

The best hypotheses are made out of a clear goal and a simple narrative. Your goal should be a meaningful and measurable metric. (It might be an existing KPI.) The narrative should be the type of campaign you are going to try. For example:

“Reminding users that a promotion is time-limited will increase transaction rate.”

Goal: increase transaction rate
Narrative: reminding users that a promotion is time-limited


“Prompting users with targeted content will increase their engagement with the site.”

Goal: increase site engagement
Narrative: prompting users with targeted content

How to come up with a good hypothesis

The best hypotheses always come from serious critical thinking. You don’t need to spend hours wading through existing data or user research. Although these can help, especially if you have no idea where to start, they can’t do the job for you.

Block out some time, either alone or with a few co-workers, and start writing down ideas. Argue about them. Play devil’s advocate. Break down the steps in your on-site customer journey. Think about the problem from its first principles. Don’t be afraid to abandon conventional wisdom.

There’s also no shortage of places to look for research and inspiration. You can check out our case studies.

You can also browse our library of modules, to get ideas flowing about the sort of things you can do with the Amigo tag on your site, and we always recommend you think about consulting some resources from the field of behavioural psychology.


You only have so much web traffic and this makes it very inefficient to try to test every idea at once. When you have a healthy list of hypotheses, it’s important to take some time to prioritise your ideas and come up with a strategy. You may want to draw up an experiment schedule.

When prioritising, you should always stack your most radical ideas towards the front. Major changes are not only more likely to deliver large, quickly identifiable uplifts, but testing these first can stop you from falling foul of one of the most common mistakes in optimisation. There is very little point optimising for colours and copy when your customer journeys have poor structure. By testing radical changes first, you will learn more and set yourself up for future success.


Good experiment design can be a challenge. If you want to split test your campaign, there are a whole host of pitfalls to watch out for.

The most important questions are ones of scale. How much traffic are you testing on? How quickly will you gather enough evidence to be sure in your conclusions? Will you get a return on your investment with a 5% uplift or a 0.5% uplift?

Last mile marketing is at its best when web traffic is segmented in the user experience as well as in analytics. This means targeting can get very precise. Even on high-traffic sites, segments can become small enough that split tests take a long time to pass. This only serves to reinforce the importance of testing bold ideas, re-engineering entire journeys, and testing radically different user experiences.


While it is important to start with a hypothesis, you don’t need to nail it down straightaway. It’s likely you will come back and revise it once you’ve spent some time specifying your campaign. Once your campaign goes live however it is very important not to change your hypothesis.

Next steps

If you have a clear hypothesis and you reckon it has a good chance of being proven right, it’s time to specify your campaign. There are two main things you will need to think about, customer journeys and customer segments.

Further reading


020 3940 4650