A, C, D, E, I, L, M, P, R, S,V, W

When we talk about personalization, we inevitably get into discussions around A/B testing and controls and variations. Of course when we make a change to our site, for whatever reason, we want a method of understanding whether that change had the desired outcome, e.g. increase in sales of a particular product line, more signups for an event, etc. This is where A/B testing comes in. It gives us a really handy tool for comparing two versions of your website: control, variation.

Further down the line, when we've got enough data to be confident in the result, we can answer questions about which version performed better against your desired outcomes. Sometimes you'll get winners, sometimes you won't, but the important point is that A/B testing gives you a much better degree of certainty that whatever changes you see are down to your efforts and not random factors.

Audiences help you show a campaign to a specific group of people that visit your site. Audiences are built from one of more conditions that are specified as attributes, such as device, location, number of conversions, and LTV, which can be used individually or combined for more specific targeting.

People that don't meet the conditions for an audience will not see your campaign. Instead they will see your default website content (this is referred to as your control).

Audience membership is evaluated each time the page loads. This ensures that your targeting is always hyper-relevant and accurate.

the average total of single orders placed on a website or mobile app over a period of time. When referring specifically to campaigns, this metric reports the average amount spent by each time an order is placed by a customer that saw the experience. AOV is important because it provides a valuable insight into how much your customers are spending on your products.

the number of orders placed by customers in the experience over a defined period of time. When referring specifically to campaigns, this metric reports the average number of days between orders on your website

much like an advertising or marketing campaign, with objectives and outcomes, a campaign is built around an idea, hypothesis, or call to action that frames a well-defined goal. That goal might be to increase sales of a product line, drive product awareness or catalog discovery, or build loyalty. A campaign is essentially a wrapper for one or more experiences, indeed one of the key advantages to campaigns is that you don't need two separate personalizations to target your different groups of users. You can create different audiences and add different content for each one all inside a single campaign. Great!

the number of times the experience link has been clicked

the percentage of visitors that clicked an experience link at least once and went on to convert

the percentage of visitors that clicked an experience link at least once

revenue from orders by visitors that clicked an experience link at least once and went on to convert

answers the question, what? Determines the message, URL, and image that are injected into your placement

your default website content without any of the changes added to an experience. For example, if you added banner content to an experience within a campaign, the visitors in the control will not see the banner

an experiment's variations, where no treatment is applied, i.e. we don't show the banner we are testing. By exposing visitors to control, we have an effective means of testing the effectiveness of your campaign

number of visitors that converted divided by the total number of visitors. When referring to side-wide impact, this metric reports the aggregated CR across your campaigns. To avoid "noise", we only look at campaigns that are live and have statistically significant results

When referring to one of your campaigns, it reports the aggregated CR across all experiences in the campaign

a simple file format used to store tabular data, such as CRM data. CSV files can be imported into the Merchandising Hub to create experiences or add audiences to an experience

the amount of money that can be attributed to the entire future relationship with each customer

an expression of how confident we are in the change reported for a given metric based on the probability of uplift and sample size

a campaign that you are still working on and hasn't yet been launched. Draft campaigns are not visible on your website and do not collect data. You'll find these campaigns in your Draft view.

the content you want your customers to see and the cohort of those customers you want to show the content to. A campaign can be built with several experiences inside it

the number of times the experience has been viewed

a prediction of the change in revenue generated by the campaign

a campaign that has been launched and . Only visitors selected randomly from your audiences will see the campaign (variation). Everyone else will see your default website content (control).

part of a site that can contain multiple placements, e.g. homepage, Global, Product Detail Page, Product Listing Page

we will report your campaign as a loser when see a less than 5% probability of an uplift and when the sample size is greater than 20%

describes the use of algorithms to train a computer to look for patterns and infer truths from data instead of being explicitly programmed to act on the data

your customers will no longer see a campaign that has been paused. Instead, they'll see your default website content. Furthermore, the campaign will no longer generate data so the reporting metrics will be frozen in time. The metrics will therefore only be valid up until the point you paused the campaign

a page that shows the details of a particular product in view and typically includes size, color, price, availability, and any other relevant information that the customer might require before making a purchase

combines a campaign, (idea), with an experience, (placement, content, and audience), to deliver content at the right time, in the right place, to the right person

a specific area on a site designated to render Merchandising Hub content, e.g. homepage Hero Banner", "Global Ribbon", "PLP Merchandising Slot". Placements are a combination of the place where they can be used, for example, your homepage and a type, for example, a banner. To help you out when choosing a placement, we've organized them into the areas in which they can be used

a page that shows a range of products belonging to a particular category or brand or the results of a search or filtering, and typically includes only essential product information such as price

a campaign can contain more than 1 experience. In fact, there really are no limits to the number of experiences you can add to a campaign, though of course, you wouldn't want to go mad! What happens when a visitor falls into both audiences? Well, we have an overlap. To get around this, you will need to define an order so we know what experience to show. You'll find a further explanation in Ordering experiences

the revenue from orders by visitors that viewed the experience and went on to convert

total revenue divided by the number of customers. When referring specifically to a campaign, this metric reports the average amount of money generated each time a visitor sees the campaign and makes a purchase

revenue per converter multiplied by conversion rate. When referring specifically to a campaign, this metric reports the average amount of money generated by each visitor that saw the campaign

the amount of data that we require to be confident that the reported change in uplift is a result of your personalization efforts as opposed to some random factor

for many of our strategies, product recommendations are built off of a seed product. This is the equivalent of the product being viewed by one of your customers. We use it in the Merchandising Hub and the Experimentation Hub to show you what products your customers would expect to see if they were viewing that (seed) product on your site

refers to the percentage of your visitors that will see the content in one of your experiences, referred to as a variation and the percentage of your visitors that will see your default website content, referred to as the control

stated simply, statistical significance means we have collected enough of a sample of the data to be confident that the observed changes in the uplift, both positive and negative, are a result of the experience being shown to your visitors and not some random factor

your website with the changes added to an experience. For example, if you added banner content to an experience within a campaign, the visitors in the variation will see the banner

the number of times an experience was viewed

the number of people that viewed the experience

when we've got enough data and we're confident that the observed change is a result of your personalization efforts and not some random factor, we can declare the campaign is a winner for a given metric