Which CRM test to run at each point of the year in eCommerce

Most brands I see test their CRM whenever they have the time. Which is, basically, never. Or in November, in full panic mode, two weeks out from Black Friday, when everyone's running around and a welcome flow that converts poorly can cost you a fortune. And the following November, same thing all over again. Right when your second abandoned cart email becomes mission-critical.

There's a calendar logic worth respecting here. Klaviyo put out a framework on this, and honestly I mostly agree with the premise. You don't test the same things in January as you do in September. Each quarter has its own window. Let me break it all down, through the lens of how we see it on the agency side.

A Cadence Calendar for Your Klaviyo AB Tests

Before we get into the quarter-by-quarter detail, two things to set up.

The first is what you test and how often. You don't launch a Klaviyo A/B test for fun. You form a hypothesis, you test it, you wait until you've got enough volume to call it, and you document the result somewhere. Without that, you'll re-run the same tests every six months without even realising it.

The second is how you isolate the effect of a test when traffic spikes. In the middle of Q4, between promos, drops and campaigns flying in every direction, you lose track of what's actually moving your numbers. So we test the heavy lifting outside the peak, and keep the peak itself purely for execution. Each quarter has a very specific role.

Q1, Validate the Flows That Run All Year

January-February is the quarter with the least commercial pressure of the entire year. Your audience just came out of your biggest revenue period, and now everything settles down. Which is exactly why you test now. You can test in a low-stakes environment. If a test flops in January, you've lost nothing. In Q4, that same failed test can torch a chunk of your year.

So Q1 is your foundation. You clean up your segmentation, you rebuild your welcome series from the ground up, and above all you put a hypothesis behind every flow that's been running on autopilot for months. What you validate in January becomes your baseline for the rest of the year.

Post-Purchase Flows, It's Now or Never

After the holidays, you've got a wave of new buyers moving through your post-purchase flows. That means more profiles passing through, so you hit statistical significance much faster. It's a point people forget. To make a clean decision on a test, you need volume. A test on a flow that three people pass through a week? You can run it six months and still learn nothing. Right now, you've got the volume. Use it.

Replenishment Flows

If you sell consumables (skincare, supplements, coffee, candles, anything that gets rebought), your holiday buyer cohort hits its first repurchase trigger in Q1. A replenishment cycle between 30 and 90 days, you test it here, because January-March is the biggest volume this flow will see all year. Test it in June and you might as well talk to a wall, nobody's passing through it.

Cross-Sell and Win-Back for New Buyers

A holiday buyer bought one specific thing. A gift, a product they got targeted on. They don't know the rest of your catalogue. Q1 is your widest window all year to teach them you also have tops, jackets, refills, whatever it is.

And you're sitting on the biggest cohort of one-time buyers you'll see all year. What I like to do here is segment cleanly between leads, one-time buyers and repeat buyers, and test win-back content built specifically to turn those one-time buyers into repeat customers. We put this in place for a cosmetics client last year and it drove a nice LTV lift on January's one-time buyers, and yet it's a project almost nobody works on.

Q2, the Quarter Where We Watched the Gap Widen

Now we're moving away from the holidays. Engagement from the profiles you collected on Black Friday starts to drop off, and depending on your vertical traffic either climbs or falls. For a home or outdoor brand, Q2 can be a big revenue period with Mother's Day and Father's Day. For others, it's quieter.

Either way, Q2 is the most under-tested quarter in eCommerce. We all want to coast until back-to-school season. Except the performance gap between brands that work their Q2 and those that slack off gets paid in full in Q3 and Q4. It's mechanical.

I see Q2 as a lab. It's the only stretch of the year long enough, calm enough and with enough traffic to validate a new flow, a new template, a new acquisition channel before the heavy half of the year kicks off.

Acquisition Offer Tests

Ad costs are lower, the promo calendar is empty, so you get super clean test conditions and fast reads. This is the moment to test your list growth offers. Percentage discount vs. free shipping vs. free sample. That tells you which incentive converts AND retains, before you scale your acquisition budget in Q3. Because the offer that fills your list fastest can also bring in your worst customers.

Evaluating AI and Predictive Tools

If you want to test a new AI feature or an optimisation tool, do it now, not when your revenue is on the line. Q2 gives you the statistical room to honestly judge whether there's lift or not. Then you decide whether you trust the thing enough to let it run in Q4.

Sunset Thresholds

We recommend starting the sunset process around 365 days of no engagement on a channel. Since most brands have a buying peak at the holidays, that means sunsetting the previous year's buyers lands right after the following year's holidays. Well played, because it gives you a big cohort of profiles to test your re-engagement strategies on. And Q2 is one of the safest quarters for that, before deliverability stakes come back in Q3 and Q4.

Mother's Day and Father's Day Gift Campaigns

These two moments are Q4 in miniature, with a lot less risk. You test your gift guide content, the gift card promo, the "shop by recipient" segmentation, and you apply the winners to your Q4 machine. You run the same mechanics as Q4, scaled down.

Q3, Lock In Everything That Compounds at Peak

Summer buying behaviour looks nothing like winter. Schedules change, people go on vacation. But more importantly, Q3 is your last full quarter to test before Black Friday eats all your team's bandwidth. Anything you haven't validated by late September, you have to wait until January. And that's a long time to fly blind through your biggest promo cycle of the year.

So in Q3, you lock in the operational details that compound at peak. Send windows, segmentation rules, deliverability hygiene, and the welcome flow that will greet the biggest acquisition wave of your year.

First up, your VIP and loyalty messaging. Your best customers cost the least and convert the fastest in Q4, so you test your VIP campaign formats, your early access and your tier-based messaging now, to tune your highest-LTV engine before the Black Friday stress test. Next, your offer ladders on cart and checkout abandonment. Abandonment volume in Q4 is massive. In the middle of Q4 you don't want to be patching up your second recovery email, so you test offer tiers, send timing and subject lines while buyer behaviour is still predictable and statistically clean. This is where getting the best time to send abandoned cart email right pays off. Your back-in-stock flow content too, because pre-holiday inventory drops generate a clear spike in replenishment subscribers. Test what converts them before Q4 demand turns that flow into a real revenue channel. And your send time, which you re-test every quarter because behaviours shift. Schedule a test in Q3, then re-block a 24-hour window to re-test it in Q4 right before your big drop. That little move saves you when the Q4 campaign crunch kicks in.

And your welcome series, again. You optimise it before the influx of new subscribers. Start simple, subject lines, CTA copy, its colour, its position, the images. Then move to the finer stuff, product block design, dynamic product blocks.

What Makes the Difference Is When You Launch Your Tests

The timing of your tests matters more than the number of them. Whether you launch five or fifteen is beside the point. What really counts is when you launch them. A brand that rolls into Black Friday with its welcome series, abandonment cadence and post-purchase flows validated months ago starts two lengths ahead of the one scrambling to test on November 15.

And above all, a test you don't track is a test you shouldn't launch. I say this to all our clients. Document every Klaviyo A/B test you run, the starting hypothesis, the result, what you took from it. Otherwise you're swapping a creative or a flow every two weeks just to say it's fresh, without ever learning a thing. You look at the previous test's results first, you pull your insights, and then you move on to the next one. A flow, like an ad, is sometimes better left running three months if it's performing rather than blown up for the sake of novelty.

To sum up, January-March you lay your foundation while it's dead quiet, in spring you place your low-stakes bets. Then at back-to-school you lock in the operational side, and in October you don't touch anything risky, you just execute.