Insights · Ramadan 1448 challenge
We announce our forecast for Ramadan 1448 before it starts, lock it, then measure it in public
The first real test of our method in a season it has not seen. We lock the spending forecast for the four weeks of Ramadan two weeks before the season and save the file’s fingerprint, so it cannot change afterwards. Then we put what SAMA published next to it, week by week, and publish the verdict even if it goes against us.
Coming soon: the forecast is locked on 24 January 2027 (16 Sha'ban 1448 AH)
The challenge and how we measure success
What we forecast
Weekly point-of-sale spending (Saudi Central Bank) in the four weeks of Ramadan, for 24 series: Saudi Arabia as a whole, the ten cities (Riyadh, Jeddah, Makkah, Madinah, Dammam, Al-Khobar, Buraidah, Abha, Tabuk, Hail), and 13 activities for Saudi Arabia. We left out education and jewelry because their historical error is too large to promise anything on them.
For each series and week: the expected figure, and the lowest and highest expected figure (a range designed to contain the actual in about 8 of every 10 weeks).
The four weeks
Ramadan 1448: 8 February – 8 March 2027 · 1–29 Ramadan 1448 AH. The week runs Sunday to Saturday as SAMA publishes it, and the four weeks are those holding most of Ramadan’s days.
- Week 1 7–13 February 2027 · 30 Sha'ban – 6 Ramadan 1448 AH
- Week 2 14–20 February 2027 · 7–13 Ramadan 1448 AH
- Week 3 21–27 February 2027 · 14–20 Ramadan 1448 AH
- Week 4 28 February – 6 March 2027 · 21–27 Ramadan 1448 AH
When we succeed
- Saudi Arabia error 8% or less: the mean gap between forecast and actual over the four weeks, relative to the actual. Outside the seasons our error is about 6%, and the season is harder.
- At least 7 of 10 cities inside the range: a city is inside the range if its actual spending falls between the lowest and highest expected figure in at least 3 of 4 weeks.
Both conditions together; otherwise we say the challenge was not met. The 13 activities are shown and measured, but do not enter the verdict.
Dates
- Forecast lock and fingerprint published: 24 January 2027 (16 Sha'ban 1448 AH)
- Method freeze: 7 February – 25 March 2027 · 30 Sha'ban – 17 Shawwal 1448 AH: no change to any forecast formula
- Every Thursday in Ramadan (11 February 2027 (4 Ramadan 1448 AH), 18 February 2027 (11 Ramadan 1448 AH), 25 February 2027 (18 Ramadan 1448 AH), 4 March 2027 (25 Ramadan 1448 AH)): we put the actual next to the forecast for every week SAMA has published by then
- The verdict: 18 March 2027 (10 Shawwal 1448 AH)
The locked forecast
Coming soon. The forecast is computed and locked on 24 January 2027 (16 Sha'ban 1448 AH) on the latest data SAMA published before it, and published here with the file’s fingerprint (sha256) and a download link, so anyone can check it has not changed since.
No figures here before the lock. What you see below is a test on last Ramadan, not a forecast for the coming one.
What it would have looked like last year
A back-test, not a real lock. We ran the same script that will lock the Ramadan 1448 forecast on Ramadan 1447 (18 February – 19 March 2026 · 1–30 Ramadan 1447 AH), after cutting the data at its notional lock day 4 February 2026 (16 Sha'ban 1447 AH): the last week the forecast saw was 18–24 January 2026 · 29 Rajab – 5 Sha'ban 1447 AH. Then we measured it against what SAMA published.
Why this test is not enough: the method itself was designed after seeing all the history, including Ramadan 1447, and the actual figures are SAMA’s latest after any revision. So the test shows the mechanism works; it does not show the method will hold in a season it has not seen. That is what Ramadan 1448 tests.
Saudi Arabia error week by week: 6.1%, 9.2%, 2.7%, 11.4%. The 13 activities for Saudi Arabia do not enter the verdict; their mean error ranges from 4.8% to 34.9%, largest in activities whose spending swings most in the seasons. City actuals fell inside the range in 28 of 40 city-weeks.
The full table: 24 series × 4 weeks
| Series | Week 122–28 February 2026 · 5–11 Ramadan 1447 AH | Week 21–7 March 2026 · 12–18 Ramadan 1447 AH | Week 38–14 March 2026 · 19–25 Ramadan 1447 AH | Week 415–21 March 2026 · 26 Ramadan – 2 Shawwal 1447 AH | Mean error | Inside the range |
|---|---|---|---|---|---|---|
| Saudi Arabia | ||||||
| Saudi Arabia (total) | 13,65412,625 – 14,767actual 14,533 · error 6.1% · inside | 14,65813,610 – 15,788actual 16,135 · error 9.2% · outside | 16,59215,289 – 18,007actual 16,149 · error 2.7% · inside | 16,48315,092 – 18,003actual 14,793 · error 11.4% · outside | 7.3% | 2 of 4 |
| The ten cities | ||||||
| Riyadh | 4,6674,311 – 5,052actual 4,863 · error 4.0% · inside | 4,9774,616 – 5,367actual 5,351 · error 7.0% · inside | 5,4685,066 – 5,903actual 5,329 · error 2.6% · inside | 5,1434,772 – 5,543actual 4,699 · error 9.5% · outside | 5.8% | 3 of 4 |
| Jeddah | 1,8931,738 – 2,062actual 1,990 · error 4.9% · inside | 2,0881,922 – 2,269actual 2,340 · error 10.8% · outside | 2,4612,265 – 2,675actual 2,421 · error 1.7% · inside | 2,3922,186 – 2,618actual 2,175 · error 10.0% · outside | 6.8% | 2 of 4 |
| Makkah | 791729 – 858actual 872 · error 9.3% · outside | 861795 – 932actual 858 · error 0.4% · inside | 1,016930 – 1,109actual 1,014 · error 0.1% · inside | 949850 – 1,059actual 886 · error 7.0% · inside | 4.2% | 3 of 4 |
| Madinah | 565510 – 625actual 615 · error 8.2% · inside | 634575 – 698actual 685 · error 7.5% · inside | 791711 – 879actual 738 · error 7.2% · inside | 822722 – 936actual 709 · error 15.9% · outside | 9.7% | 3 of 4 |
| Dammam | 683624 – 748actual 689 · error 0.8% · inside | 706648 – 769actual 744 · error 5.1% · inside | 773707 – 846actual 734 · error 5.3% · inside | 726665 – 792actual 671 · error 8.2% · inside | 4.9% | 4 of 4 |
| Al-Khobar | 378348 – 410actual 380 · error 0.5% · inside | 394364 – 427actual 411 · error 4.0% · inside | 431397 – 467actual 408 · error 5.5% · inside | 432399 – 468actual 396 · error 9.0% · outside | 4.7% | 3 of 4 |
| Buraidah | 327302 – 353actual 371 · error 11.8% · outside | 361335 – 389actual 414 · error 12.6% · outside | 410378 – 446actual 413 · error 0.7% · inside | 420382 – 461actual 376 · error 11.7% · outside | 9.2% | 1 of 4 |
| Abha | 144119 – 176actual 153 · error 5.6% · inside | 156128 – 190actual 176 · error 11.4% · inside | 172140 – 211actual 169 · error 2.1% · inside | 181148 – 221actual 152 · error 18.7% · inside | 9.4% | 4 of 4 |
| Tabuk | 241211 – 274actual 255 · error 5.5% · inside | 257227 – 290actual 286 · error 10.1% · inside | 270239 – 306actual 256 · error 5.7% · inside | 263233 – 297actual 220 · error 19.6% · outside | 10.2% | 3 of 4 |
| Hail | 203181 – 228actual 226 · error 10.1% · inside | 222199 – 248actual 254 · error 12.5% · outside | 253224 – 285actual 240 · error 5.2% · inside | 273239 – 311actual 226 · error 20.3% · outside | 12.0% | 2 of 4 |
| 13 activities for Saudi Arabia | ||||||
| Restaurants & cafés | 1,2631,151 – 1,386actual 1,217 · error 3.8% · inside | 1,4161,295 – 1,548actual 1,396 · error 1.5% · inside | 1,6671,522 – 1,825actual 1,392 · error 19.7% · outside | 2,1761,956 – 2,421actual 1,602 · error 35.9% · outside | 15.2% | 2 of 4 |
| Food & Beverages | 1,9871,795 – 2,199actual 2,326 · error 14.6% · outside | 1,8631,688 – 2,057actual 2,500 · error 25.5% · outside | 1,8821,704 – 2,080actual 2,257 · error 16.6% · outside | 1,9391,763 – 2,133actual 2,114 · error 8.3% · inside | 16.2% | 1 of 4 |
| Apparel, Clothing & Accessories | 1,7621,494 – 2,078actual 1,899 · error 7.2% · inside | 2,3261,972 – 2,743actual 2,502 · error 7.0% · inside | 3,0352,508 – 3,672actual 2,833 · error 7.1% · inside | 3,0822,471 – 3,843actual 2,334 · error 32.0% · outside | 13.3% | 3 of 4 |
| Health | 845762 – 938actual 900 · error 6.0% · inside | 902810 – 1,005actual 972 · error 7.2% · inside | 963862 – 1,075actual 946 · error 1.7% · inside | 775693 – 866actual 824 · error 6.0% · inside | 5.2% | 4 of 4 |
| Gas Stations | 917866 – 970actual 875 · error 4.7% · inside | 937888 – 989actual 898 · error 4.4% · inside | 952893 – 1,015actual 840 · error 13.3% · outside | 999931 – 1,072actual 825 · error 21.2% · outside | 10.9% | 2 of 4 |
| Electronic & Electric Devices | 180155 – 210actual 201 · error 10.2% · inside | 171147 – 200actual 207 · error 17.0% · outside | 186159 – 217actual 195 · error 4.9% · inside | 143123 – 168actual 152 · error 5.6% · inside | 9.4% | 3 of 4 |
| Hotels | 344286 – 414actual 376 · error 8.4% · inside | 414349 – 491actual 335 · error 23.7% · outside | 501427 – 588actual 450 · error 11.4% · inside | 441365 – 534actual 362 · error 21.8% · outside | 16.3% | 2 of 4 |
| Construction & Building Materials | 355332 – 380actual 378 · error 6.1% · inside | 349323 – 376actual 389 · error 10.3% · outside | 366338 – 396actual 371 · error 1.4% · inside | 243223 – 264actual 247 · error 1.6% · inside | 4.8% | 3 of 4 |
| Furniture & Home Supplies | 490402 – 598actual 618 · error 20.7% · outside | 513403 – 654actual 607 · error 15.4% · inside | 595453 – 780actual 605 · error 1.7% · inside | 492364 – 665actual 458 · error 7.4% · inside | 11.3% | 3 of 4 |
| Recreation & Culture | 361290 – 450actual 323 · error 11.7% · inside | 390309 – 494actual 339 · error 15.3% · inside | 487382 – 622actual 331 · error 47.0% · outside | 562432 – 731actual 339 · error 65.7% · outside | 34.9% | 2 of 4 |
| Telecommunication | 200175 – 229actual 196 · error 1.8% · inside | 203175 – 234actual 213 · error 4.8% · inside | 208180 – 242actual 189 · error 10.3% · inside | 195167 – 226actual 174 · error 12.1% · inside | 7.2% | 4 of 4 |
| Transportation | 1,058922 – 1,213actual 1,167 · error 9.4% · inside | 1,144977 – 1,339actual 1,129 · error 1.3% · inside | 1,2021,017 – 1,422actual 1,081 · error 11.3% · inside | 833696 – 996actual 777 · error 7.1% · inside | 7.3% | 4 of 4 |
| Public Utilities & Services | 5746 – 70actual 60 · error 5.7% · inside | 5141 – 64actual 63 · error 18.4% · inside | 5141 – 64actual 58 · error 11.8% · inside | 4436 – 55actual 45 · error 0.6% · inside | 9.1% | 4 of 4 |
Fingerprint of the test file (sha256): 3487501da157ed03306fd55600e3bedeb652ab20144f31c32f8fffc18cb76723 · Download the file
The method and its limits
- Method: the hybrid published in the methodology: outside the seasons, the same week last year multiplied by growth over the last 4 weeks; inside the seasons, the same season in the two previous Hijri years. There are several weeks between the latest published data and Ramadan, so each forecast week feeds into the next.
- Lowest and highest expected figure: from the same method’s errors the last 26 times it forecast the same distance ahead. These errors are mostly from outside the seasons, so the range may be too narrow for Ramadan.
- What we do not measure: card spending at point of sale only, before removing price inflation, for a whole city or activity, not a single shop. It excludes cash and transfers.
- Script and files:
scripts/ramadan_lock.pygenerates the file and does not write over an existing one; every figure on this page is recalculated automatically at each build, and publishing stops if any differs. - Source: Saudi Central Bank (SAMA), weekly point of sale; the calculation is DataPulse’s and is not an official statistic. Analysis to support decisions.