A raised bed that bows is usually the rails
I looked at the veg bed in March after a wet winter. The long side had a belly in it. The top rail had come off the corner post by a finger's width. I thought the timber had warped in the sun. It was February rain in the soil. The rails were too thin, and the only screws were into the end of the board. Wet soil had pushed. The middle bowed. The corner let go.
A raised bed is a box that holds a volume of soil off the path. The rails are the long sides of that box. They have to stay straight under a winter of wet earth. A compost bin is a box, not a bag of compost is the same idea in another corner of the plot. If the sides give, you do not have a bed. You have a slump with a timber necklace.
The soil is the load. Dry soil is heavy. Wet UK soil is heavier. A long, thin rail with no post in the middle is a beam with a sack hanging on it. It will belly. Corners take the other half of the job. If the board is only screwed into end grain, the screws walk. You see a gap at the post first, then a curve along the side. People blame the timber. The timber was doing a rail's job without a rail's thickness, and without a fixing that could hold.
When I check a bed I sight along the long sides. If I can see a curve, I push the middle with my palm. Movement means there is no backing. Then I look at the corners. I want a post or a bracket, and screws that go into the post, not into the end of the board and hope. Rusted screws that turn to dust are a corner already open. I also look at the soil line on the inside face. Gravel boards rot at the soil line is that wet on a fence. A rail sitting in the same damp, with soil piled over the top edge, is rotting while it bows. I scrape the soil back and press a spike into the timber at ground level. If it sinks, the rail is finished. If it only marks, I still want a mid-span stake before the next winter.
A fork and a spade are not the same digging tool. I would not lever a stuck clump against the rail. That is how you finish a bow that had only started. Fork the soil, lift it, do not use the side as a fulcrum. If I am emptying a bed to reset a side, the fork comes out through a gap I make on purpose, not through a hatch I invent by standing on the timber.
Path edging that lifts is usually the frost is a different shove. Frost levers a shallow line. A bed bows because the rails cannot hold the wet load. I would not treat a raised bed as a deep piece of path edging. If the sides are only a single course of bricks with no return, frost will lift them and wet soil will still kick them out. That is two failures on one cheap wall.
Skip the kit of thin slats and plastic corners. The corners crack. The slats cup. You get a year. Skip filling a new tall bed in November and walking away. You have built a water butt with timber walls. Skip lining that holds a lake against the inside face. A liner with no drainage is a wet bag in a rotting box. Skip standing on the rail to reach the middle. I have done that. The belly started there. Skip using one long board with no mid post on a bed longer than a stretch of my arms. If I cannot reach the middle without leaning on the side, the side needs a stake.
This shows after a wet winter, usually March, when you first look at the plot properly. Autumn is when I would put a mid post in, drop the soil a little below the top rail, and change any screw that has already stained the corner. I walk the beds after a long wet week, not on a date from a catalogue. If the side moves under my hand, it will move more by February.
I do not need four new beds because one rail bowed. I need sides that stay straight when the soil is wet.
If I later add a buy link here, it will be a paid promo.