Scott Services: Best Tweets from #SEMrushchat: How to Run Social Media Experiments

There is no absolute how-to guide for social media marketing because every business is different, every audience is diverse, each audience’s experience is different, and platform algorithms often make organic reach difficult.

Every marketing plan has different goals — conversions, engagement, brand awareness, etc. — and every goal needs a different strategy aimed at a particular audience. Every share needs a different message, graphic/video, and landing page to meet the needs of your target audience. Social media marketing can be challenging and requires a lot of planning, time, and effort.

You can plan well and use your time efficiently, and still not get the results you need or want. Experimentation is often necessary to find strategies that will work and meet your goals. So we asked our community and guest Mike Allton during SEMrushchat about testing, factors to consider, bad results, steps for running an experiment, and data to consider for ideation. Below you will find some of the best tweets from the chat.

Note: You can retweet any of the tips/insights below by clicking on the Twitter logo next to the quote.

Yes. Here is why: Too many so-called experts will talk about best practices. But the real experts will give the caveat that social media is not a one size fits all approach. There are a lot of considerations and moving parts. Test to see what works.

Testing on social media should be part of your strategy and everyday tactics. Why? What is working today may not work tomorrow. Our audience expectations change by the day & it is best practice to have continuous optimization, and the way to that is testing.

Testing is a must on any platform of marketing. Your audience, trends, intent, tech, etc. are constantly changing and/or updating. Testing helps you stay on top of what is happening while knowing which elements do or don’t work for you

Not trying to be contrary here… but. No. Testing is not always a must. – if your social media is more about relationship building or human engagement, testing isn’t always a must. – if your social media is more transactional, yes.

100% – testing is a must! What works in one vertical could have almost no effect on another. Without testing, you lack the data needed to make informed decisions

If I had my way every time, yes, yes, it would. But that’s not how agency-life works sometimes. Running a live campaign and judging the results in and of itself is a form of experimenting, too.

Platform (i.e., Twitter). Tactics (i.e., Tweet Chats). Medium (Text/Images/Video). Style (i.e., Funny or Professional). Paid Promotion. Timing & Frequency.

1) Graphics presentation 2) Colour patterns 3) Post text type 4) Post timing 5) Brand position

There are so many ways to test on social media: Different images, audience targeting for paid social, posting times, new platforms/tactics, shorter or longer copy, sending traffic to different landing pages and so much more. I testing and innovation!

Content/copy, placements (right side vs. in-message), traffic vs. conversion campaigns, click types, paying for impr. vs. clicks. There’s a lot of levers to experiment.

Considerations for social media testing: Duration of engagement, short term or longer due to amplifications. Sentiment towards content (comments can be negative)! Branded hashtag reception.

Just a few things worth testing are: Posting time to maximize engagement, post frequency, CTA vs. evergreen frequency, images to improve engagement, retargeting funnels, engagement on particular platforms, copy, and phrasing.

Did a contest on FB for healthcare staffing. To get a new audience, we used an ad instead of organic posting (FB was their best social platform). It bombed, and we got 1 entry. Learned that organic posting can still perform better than ads.

I think it has to be during my initial years of learning paid Marketing. Not using the right target audience cost me the engagement and also burnt money in the wrong place. Learned how to optimize it better from there on.

One of my favorites was an ecomm company that sold 1 product on their website. The CTR was horrible, and the CPC was close to what they expected the CPA to be. Stopped running ads to their site and straight to their Amazon store, and both improved almost instantly.

I’ll be brutally honest; I managed a campaign (organic and paid) that was a complete flop. The targeting sucked, budgets were set too low considering the competition, and copy didn’t resonate with the audience. Lesson learned… never do that again! Plan better.

I’d consider a social media test as a failure if it doesn’t provide an answer to a question. When that happens, it’s more a process failure than a social media failure.

Sometimes focusing on expanding your audience is the wrong move. Connecting with existing fans/audiences will keep them happy/engaged and bring in other like-minded individuals. Don’t expand too rapidly and lose the engagement you’ve already developed.

First, understand that social media tests are both easy and impossible at the same time. It’s relatively easy to set up and execute a test, but technically they’re not controlled enough to be scientific, so beware.

1: Set Goals 2: Prioritize Goals 3: Design your Experiment 4: Test Ideas 5: Analyze and Learn 6: Repeat.

Evaluate goals. Establish a hypothesis. Choose KPIs for measurement. Create content. Schedule content. Let it gather data. Look at results (from KPIs) Implement what works. Learn from what doesn’t work.

1. Find a question you want answered. 2. Change one aspect of the post (whatever) and wait. 3. See whether it worked and answered your question. 4. Yes? Well done! 5. No? Change a different aspect.

First steps should be identifying goals, what you want to see, and determining the variables that will be tested between the different posts, i.e., picture vs. no picture. You should also figure out how you’ll measure its success, such as likes, comments, or leads.

Trending topics, influencers, past experiments, and recent success are all factors which should inspire your ideas for your next social experiments!

Data that inspires us to test new ideas include: Spikes in engagement on certain types of posts. Our competitors and audience using a certain hashtag. A significant drop in reach for certain content/platforms. A day/time that outperforms others.

Lead gen and revenue attribution, ultimately. Fun, trendy social content is cute, but if it’s not boosting the bottom line, it is a waste of time.

As a marketer, I often find that I’m too afraid to just try things. We tend to be hyperfocused on theory and data pre-testing. Really. We’re so insecure and boring. Young/social/party/media/music type people can inspire and push you. So get involved with them.

Factors that I look at are around the audience or customers. How are we best communicating and reaching the audience, and how are they engaging with our content. Usually, it starts with what is working and what isn’t and how can we improve engagement or conversion rates.

Test inspiration -> High performing concepts in your vertical. High performing concepts outside your vertical. High performing content on your blog.Highly trafficked content on competitor sites. Look at engagement metrics, BUT if you own the channel, look at multi-channel funnels in Google Analytics!