I Fixed My Sleep Before I Fixed My Testosterone – Here’s Why That Was the Right Order

My first instinct was to optimize testosterone. Buy the supplements, read the protocol, run the bloodwork and target the number. Ron Males’ first question was: “What does your Oura say about your deep sleep?”

I almost dismissed it. I came to him with a T question. He answered with a sleep question. It took me about two weeks of reading to understand that those aren’t different questions.

TL;DR: I fixed sleep before touching any testosterone-specific protocol. T went from 412 to 638 ng/dL over six months. IGF-1 went from 118 to 224 ng/mL. The sleep work drove the hormonal work — not the other way around. Every T supplement I’ve tried since then has moved bloodwork less than fixing deep sleep did first.

Why Sleep Is the Testosterone Question

The reason Ron asked about sleep first isn’t because sleep “helps” testosterone. It’s because sleep IS testosterone production, for the largest portion of the daily cycle.

Somewhere between 60–70% of daily testosterone is synthesized during sleep — specifically during slow-wave sleep (stage 3) and the testosterone pulses that follow REM cycles. The LH signals that tell the testes to produce T are sent primarily during sleep, timed to the sleep architecture. Compress or fragment sleep, and you’re directly cutting production time.

IGF-1 has an even stronger sleep dependence. The largest single GH pulse in any 24-hour period happens during the first slow-wave sleep cycle — typically 60–90 minutes after sleep onset. That pulse is responsible for the majority of IGF-1 synthesis. My first flagship tracking article (the one that documented IGF-1 and deep sleep correlation over six months) made this point with four checkpoints of data. The mechanism is real and it scales: more deep sleep, more GH pulse, more IGF-1.

When Ron said “sleep first,” what he was actually saying was: “You’re trying to optimize an engine that’s only running half its cylinders. Fix the cylinders before you touch the fuel injectors.”

Baseline — Month 0 (Before Anything Changed)

Marker	Month 0	Units	Notes
Deep sleep (30-day Oura avg)	38	min/night	Pathetically low. Toddler was 18 months.
Sleep efficiency (Oura)	71	%	Benchmark: 85%+ is good
HRV (30-day avg)	42	ms	Suppressed for my age range
Total testosterone	412	ng/dL	Morning fasted draw
Free testosterone	10.2	pg/mL	Low-normal
IGF-1	118	ng/mL	Bottom 20th percentile for my age
Sleep duration avg	6.1	hours	On paper acceptable. In practice, poor architecture.

The sleep duration number is deceptive. 6.1 hours looks close to adequate. The deep sleep number — 38 minutes — is the real story. (Sleep efficiency is the percentage of time in bed actually spent sleeping rather than awake — below 80% suggests fragmented sleep even if total time looks acceptable) Sleep efficiency at 71% meant I was spending significant time in bed not in meaningful sleep stages. Between the toddler’s overnight waking (which was happening 3-4 times some nights during the worst months), my own phone use, and inconsistent bedtimes, my sleep architecture was completely chaotic despite spending 7-8 hours in bed on many nights.

What Changed First — No Supplements in Month 1

Ron’s rule in the Anabolic Alchemy framework is explicit: behavioral interventions precede supplements. Not because supplements are bad, but because supplements layered onto a broken sleep system produce worse results than behavioral intervention alone. He’s seen men run high-dose vitamin D and zinc for months while sleeping five hours a night and wondering why bloodwork isn’t moving.

Month 1 protocol — behavioral only:

• Fixed wake time: 6:30 AM every day including weekends. The circadian anchor.
• Bedroom temperature: 65°F. I bought a standalone AC unit for the bedroom. Best $180 I spent that year.
• Morning sunlight within 20 minutes of waking: 10-minute walk outside, no sunglasses. This anchors the (suprachiasmatic nucleus, the brain’s master circadian clock, which is set primarily by light exposure in the morning — anchoring it produces more consistent cortisol awakening response and better sleep-wake timing throughout the day) SCN and produces the cortisol awakening response that healthy sleep architecture depends on.
• No caffeine after noon. My previous cutoff was 2-3 PM. The 5-6 hour caffeine half-life means 2 PM caffeine is still partly active at 10 PM bedtime.
• Phone physically out of the bedroom. Not silent — out of the room.

Nothing bought. Nothing subscribed to. Just consistent execution of the behavioral variables that actually set the circadian system.

Month 2 Check-In

Marker	Month 0	Month 2	Change
Deep sleep avg	38 min	68 min	+30 (+79%)
Sleep efficiency	71%	82%	+11 points
HRV avg	42 ms	48 ms	+6 (+14%)
Total testosterone	412 ng/dL	482 ng/dL	+70 (+17%)
IGF-1	118 ng/mL	148 ng/mL	+30 (+25%)

Two months of behavioral sleep work, no supplements, no T-specific protocol. T up 70 points. IGF-1 up 30 points. I hadn’t touched anything specific to testosterone production.

Deep sleep doubled from 38 to 68 minutes. This is where the mechanism shows up in the data: every 10 additional minutes of slow-wave sleep is another 10 minutes of GH pulsatility. The IGF-1 gain at month 2 is directly traceable to the sleep architecture improvement rather than any supplementation, because I hadn’t supplemented anything yet.

At the month 2 mark, Ron added magnesium glycinate. 400mg before bed. The rationale: magnesium directly supports slow-wave sleep induction by promoting GABA activity and parasympathetic tone. It’s not sedating in the pharmaceutical sense — it makes it easier for the nervous system to transition into deep sleep rather than fragmenting into light sleep and partial waking. The full story on why form matters is in the magnesium glycinate vs. oxide piece — short version: the glycinate chelate crosses the blood-brain barrier more effectively than oxide or citrate forms.

Month 4 Check-In

Marker	Month 0	Month 2	Month 4	Change (0→4)
Deep sleep avg	38 min	68 min	88 min	+50 (+132%)
Sleep efficiency	71%	82%	88%	+17 points
HRV avg	42 ms	48 ms	54 ms	+12 (+29%)
Total testosterone	412 ng/dL	482 ng/dL	548 ng/dL	+136 (+33%)
IGF-1	118 ng/mL	148 ng/mL	182 ng/mL	+64 (+54%)

At month 4, Ron added a basic supplement stack — zinc picolinate 30mg, vitamin D3 5000 IU + K2 200mcg, boron citrate 9mg. These were addressing deficiencies the bloodwork confirmed, not optimizing on top of sufficiency. My D3 was at 24 ng/mL (low). Zinc wasn’t tested but my dietary pattern suggested deficiency.

The T trajectory continued moving — from 482 at month 2 to 548 at month 4. But notice that this move (66 points in 8 weeks) happened at the same rate as the first two months even though I added the supplement stack at month 3. The supplements contributed to the continued progress, but the sleep foundation was doing most of the work. The stack didn’t rescue a broken sleep environment. The sleep environment was already working, and the stack provided marginal support on top of it.

This is Ron’s point about sequential optimization, and it took me until month 4 to fully internalize it: the supplement stack running on top of my month-4 sleep is producing far better results than the same supplement stack would have produced on top of my month-0 sleep. Sequence matters as much as the interventions themselves.

Month 6 — Final Results

Marker	Month 0	Month 6	Total Change
Deep sleep avg	38 min	102 min	+64 (+168%)
Sleep efficiency	71%	91%	+20 points
HRV avg	42 ms	58 ms	+16 (+38%)
Total testosterone	412 ng/dL	638 ng/dL	+226 (+55%)
Free testosterone	10.2 pg/mL	18.6 pg/mL	+8.4 (+82%)
IGF-1	118 ng/mL	224 ng/mL	+106 (+90%)

55% total T improvement and 90% IGF-1 improvement over six months, starting with eight weeks of behavioral sleep intervention and no supplements. The correlation between deep sleep minutes and the hormonal markers tracked in near-lockstep across every checkpoint. That wasn’t accidental.

The morning sunlight piece covers one of the single most important behavioral interventions in detail — the circadian anchoring effect of outdoor light within the first hour of waking produces downstream sleep architecture benefits that compound over weeks. My deep sleep improvement from month 0 to month 2 was almost entirely attributable to two changes: consistent wake time and morning sunlight. Everything else built on that foundation.

The Sequential Optimization Framework

What I’d Do If Starting From Scratch

Tier 1 (free, do these first and give them 6-8 weeks):

• Fixed wake time, same daily including weekends
• Outdoor light exposure within 20 minutes of waking (10-15 minutes minimum)
• Bedroom at 65-67°F
• Caffeine cutoff at noon, hard stop
• Phone out of the bedroom entirely

Tier 2 (add after Tier 1 is 6+ weeks consistent):

• Magnesium glycinate 300-400mg before bed
• Blue light blocking glasses starting 2 hours before intended sleep
• Consistent bedtime (not just wake time)

Tier 3 (add once sleep efficiency is consistently above 85% on Oura):

• Vitamin D3 + K2 if deficient
• Zinc if deficient
• Glycine 3g before bed (additional slow-wave sleep support)
• Apigenin 50mg if sleep latency is still elevated

Tier 4 (only if Tier 3 is locked and sleep is consistently good):

• T-specific supplements — boron, tongkat ali, ashwagandha
• Training optimization for hormonal response
• Bloodwork-guided adjustments

Notice where the T-specific work falls: last. Not because it doesn’t matter, but because Tiers 1-3 produce more hormonal movement than Tier 4 will produce without them.

Things My Friends All Wanted to Know

What if my sleep is broken because of kids? Mine was. The toddler waking was a real variable. Ron’s advice: control what you can control. Consistent wake time even on nights with waking. Morning sunlight regardless of how rough the night was. The nights the toddler sleeps through, the architecture will be there to take advantage of. Don’t let the uncontrollable nights become a reason to abandon the controllable variables.

Can I track sleep accurately with Oura? Not as precisely as clinical polysomnography. But the trends within your own data are reliable for decision-making. I’m not comparing my Oura numbers to anyone else’s. I’m tracking my own trend over weeks. The direction of that trend is what matters.

How long before sleep changes produce hormone changes? In my data, T started moving within 6-8 weeks of consistent behavioral sleep improvement. IGF-1 moved similarly. These aren’t instant changes — the hormonal system operates on weeks-to-months timescales. Eight weeks of consistent execution is the minimum before pulling bloodwork to assess.

Will magnesium make me drowsy? Not in the sedating way sleep drugs do. Magnesium glycinate at the doses I used promoted easier transition into slow-wave sleep without morning grogginess. The full breakdown of what magnesium glycinate did specifically to my sleep, HRV, and T is in a separate piece — the short version is that it added roughly 6-8 minutes of additional deep sleep on average in my Oura data when added at month 2.

The Right Order

Six months later: deep sleep tripled, T up 55%, IGF-1 up 90%. I’d almost gone down the peptide rabbit hole at month 0. The fact that I didn’t — the fact that Ron redirected me to sleep before anything else — is the reason the numbers moved as much as they did. Peptides layer on top of a working system. They don’t repair a broken one.

Fix sleep first. Everything else follows.