Scaling SMS campaigns with split testing

‍

Split testing is an important part of scaling SMS marketing campaigns. In SMSEdge we work with our clients to split test every aspect of the campaign to truly hone in on maximizing revenue. 

First thing first, to conduct a fruitful split test you must be able to track the performance of the SMS campaign. You will need to track how many SMS were sent, how many users clicked on the link, signed up, and/or purchased your product.

‍

‍

Since the number of people who opened the SMS is a major factor in the rest of our measurements, it’s crucial that we’ll track our SMS campaign’s click rates correctly to get the most accurate numbers.

How to track SMS link clicks?

To do that, each SMS will need to be sent with a unique link.

For example -

SMS 1 - Hi Jeff, we’re still waiting for you at http://smse.dg/a

SMS 2 - Hi Amy, we’re still waiting for you at http://smse.dg/b

SMS 3 - Hi Trevor, we’re still waiting for you at http://smse.dg/c

When Amy will open the SMS link we’ll know that she’s the one who opened it and not Trevor or Jeff, and we’ll count it as one click.

What happens when Amy opens the link twice?

So now we sent 3 SMS’s and got 1 click from Amy. If Amy opened the link twice, then we get two clicks right? Well not exactly. 

At SMSEdge we ignore additional clicks of customers in the campaign to maintain a more accurate view.

This way, if we have two clicks we know that both Amy and Jeff (or Trevor) clicked, and it wasn’t just Amy clicking twice.

What happens if we get too many clicks, link previews, or desktop clicks?

In that case, you will need to compile a list and filter it by the different browser user agents and IPs.  

When you are using SMSEdge, the platform automatically does that for you. Therefore, the clicks you see are real, unique clicks.

‍

‍

‍

‍

‍

There are a few ways to track signups and sales conversions, the most commonly used is the server to server pixel integration.
In SMSEdge you have the option to integrate a dynamic parameter that is generated by the platform and passed to the user’s landing page when he opens the link. 

It can be done by adding the {{{dynamic}}} parameter in the offers tab of the campaign.

Then when Amy opens this link www.example.com/?d={{{dynamic}}} it’ll change to www.example.com/?d=2038174 this 2038174 is a unique number for that SMS. Later, once a signup or a sale is made it is passed back to the server as:

‍

This ability to track and evaluate the performance creates the perfect conditions for an optimal split test. 

Now we will show you how you can split test ad texts, landing pages, routes and more using our platform.

‍

‍

‍

‍

Real time performance tracking in SMS marketing

To start we’ll load the different variations of our marketing materials into the SMS campaign builder. Then, as 94% of SMS are opened within the first 3 minutes of receiving the SMS, it’s wise to send a small batch, and based on the results of your experiment, you will choose the best performing marketing materials and continue sending the campaign.

Successful split testing strategies

The strategy depends on the country and the offer type that you’re planning to send. Generally speaking we recommend starting with about 200 SMSs as an initial “test batch” per SMS communication route. Although this test size is not statistically significant, we can eliminate outliers like SMS texts that are getting blocked, landing pages that are not converting, or routes that are disconnected. As the campaign progresses, we will gain more data to make well informed decisions.

In SMSEdge, the platform will make the first batch calculation according to:

Batch size (200 SMS) X number of routes X number of lists

‍

LANDING PAGES

To convert your clicks into signups you will need to test different landing pages and optimise according to the results. The landing pages need to be adjusted to mobile and have the offer that converts and resonates with your client. 

On some pages the conversion from a click to signup can take up to 30 minutes, but on most pages it will take less than 5 minutes. 

As landing pages and offers might have a different sales value (whether CPA - Cost per acquisition, CPL - cost per lead or CPC) you need to take that into account when measuring the campaign results.

The usual suspects for an underperforming pages are

1. The message is different from what they thought they’ll receive in the SMS
2. They don’t understand the landing page content
3. The visuals aren’t talking to them
4. They don’t like or want to buy the product
5. There is a technical issue like the page not working well on mobile

Another sure way is to get someone from your target audience to check the page.

Technically monitoring what the user is doing on the page can be done with tools like Hotjar.

ADTEXTS

Each campaign should contain at least 3 to 4 variations of ad text. This is one of the best ways to learn about your customers. We suggest using different ad texts to track the response rate. For example, an ad can have the most clicks but the fewest conversions and vice versa. 

After a test send, you will be able to choose the best performing ads to send to the rest of your contacts.

How different should the SMS texts be?

We recommend putting one specific SMS text that describes your offer, another SMS text that is generic to create curiosity (Also known as click bait), and one more if you can think of something creative.

So remember - Generic, specific, creative.

ROUTES

‍

‍

With the ability to test the performance of routes you get an option to significantly discount your SMS rates and continue to receive the best delivery. This is especially important for SMS marketers as sending for example a 1,000,000 SMS campaign at a rate of Є/$ 0.03 campaign would eventually cost Є/$ 30,000 . Thus maintaining an optimal price per SMS is crucial to receiving a positive ROI.

The ability to choose routes is unique to SMSEdge. While in most SMS marketing platforms you can only send through one route, in SMSEdge you can select a few routes for the country and compare the performance of the routes to receive optimal pricing and deliverability ratio to your content. 

We recommend selecting up to 3 routes depending on your list size and selecting first the routes that worked well for you before.

‍

Depending on your vertical, the mobile carriers of your contact numbers, and more factors, the routes’ performance can vary. For these reasons split testing an SMS campaign is important.  

‍

‍

‍

By now you’re probably wondering how does all this come together, we go

1. Create campaign.
   

2. Upload a list.
   

3. Add the offers - We recommend adding at least 2 landing pages at first to see their variance.

‍

4. Add the SMS texts (ad texts) - We recommend starting with 3 to 4 ad texts.

‍

      5. Selecting the routes - Depending on your list size we recommend starting at between 3 to 6 routes and narrowing it down to 1 once the first.            batches were sent.

6. Sending in batches - Send the first batch (approx. 400 SMSs) and wait for the first clicks, then we recommend selecting the routes that performed best and the ad texts that had been opened.

‍

‍

Sign up today to SMSEdge to start split testing and optimizing your campaign!