A live dashboard doesn't save you money by itself — someone still has to look at it and act. These five things, done in your first week, are what turn "we have visibility now" into an actual reduced bill by the time your next invoice arrives.
1. Tag your top features first — everything else is easier once spend is attributed to something specific.
2. Set a rough budget alert immediately; refine it later.
3. Review and act on your first recommendation before the week is out — that's the step most teams skip, and it's the one that actually moves the bill.
The Five Things
Tag your top 3-5 features
~15 minutes
Without feature tagging, your dashboard shows one undifferentiated number. With it, cost breaks down by product area, and "the bill went up" becomes "the summarization feature's cost went up" — a diagnosis instead of a mystery. Start with your three or four highest-volume features; you can tag more later, but the highest-impact ones are worth doing on day one.
Set a budget alert — even a rough one
~5 minutes
Don't wait for a perfectly calibrated number. Set a threshold slightly above what your first few days of observed spend suggest, and tighten it as you get more data. An imperfect alert set today catches a runaway feature or a leaked key days sooner than a perfect alert set next month.
Review your first recommendations — and act on one
~20 minutes
This is the step that actually moves your bill, and it's the one most teams put off. Once your first recommendation is ready (typically within 24 hours), open it, read the specific pattern it's based on, and implement at least the highest-impact one before the week ends. Visibility without action doesn't save anything — this is where the savings actually happen.
Ready to work through these?
All five live in your dashboard — tagging, budgets, and recommendations in one place.
Open Your DashboardTag your top 3 features. Set a budget alert. Review recommendations.
Invite a teammate
~2 minutes
Cost visibility that lives in one person's head or one person's login doesn't become a team habit. Adding at least one other engineer — whoever owns the highest-cost feature you tagged in step one — turns this from a solo side project into shared context that survives you being on vacation or switching teams.
Check your cache hit rate
~5 minutes
If you're not already caching duplicate or near-duplicate requests, this number will likely surprise you — production traffic commonly has 15-30% duplicate or near-duplicate requests that a caching layer would catch for free. A low or zero hit rate in week one is a flag, not a failure; it's telling you where the next easy win is.
What This Week Sets Up
None of these five actions individually takes more than 20 minutes, and together they take under an hour spread across the week. What they set up is bigger than the time investment suggests: attributed cost data, a safety net against runaway spend, at least one implemented saving, a second person who can act if you're unavailable, and visibility into your caching opportunity. That's the foundation everything else — an ongoing weekly review, a wider team rollout — builds on.
Frequently Asked Questions
What should I do first after connecting Preto?
How do I tag a feature in Preto?
Should I set a budget alert immediately, or wait until I understand my baseline?
Do I need a team account to invite a teammate?
Work through your first-week checklist now.
Tagging, budget alerts, recommendations, and cache hit rate — all live in your dashboard, ready when you are.
Open Your DashboardQuestions along the way? Reply to your welcome email.
